index.html

   1 <html>
   2         <head>
   3                 <title>Dscho's git log</title>
   4                 <meta http-equiv="Content-Type"
   5                         content="text/html; charset=UTF-8"/>
   6         </head>
   7         <body style="width:800px;background-image:url(dscho.git?a=blob_plain;hb=832be85c785c80202f17b87db7f063ae57ec2cac;f=paper.jpg);background-repeat:repeat-y;background-attachment:scroll;padding:0px;">
   8                 <div style="width:610px;margin-left:120px;margin-top:50px;align:left;vertical-align:top;">
   9                         <h1>Dscho's git log</h1>
  10                         <div style="position:absolute;top:50px;left:810px;width=400px">
  11                         <table width=400px bgcolor=#e0e0e0 border=1>
  12                         <tr><th>Table of contents:</th></tr>
  13                         <tr><td>
  14                         <p><ul>
  15                         <li><a href=#1236554268>09 Mar 2009 So, what is missing from my <i>rebase-i-p</i> branch?</a>
  16                         <li><a href=#1236479389>08 Mar 2009 New Git for Windows version</a>
  17                         <li><a href=#1235092615>20 Feb 2009 Code reviews</a>
  18                         <li><a href=#1234409395>12 Feb 2009 Interactive <i>rebase</i> just learnt a new command: <i>topic</i></a>
  19                         <li><a href=#1234320806>11 Feb 2009 Thunderbird, oh Thunderbird, you always make my small brain hurt</a>
  20                         <li><a href=#1234141489>09 Feb 2009 <i>format-patch --thread</i> and Alpine</a>
  21                         <li><a href=#1234140696>09 Feb 2009 <i>rebase</i> updates</a>
  22                         <li><a href=#1234040744>07 Feb 2009 The infamous <i>mark</i> command in the <i>rebase</i> command</a>
  23                         <li><a href=#1233707628>04 Feb 2009 New valgrind series</a>
  24                         <li><a href=#1233706294>04 Feb 2009 Problems with split-topic-branches.sh</a>
  25                         </ul></p>
  26                         <a href=dscho.git?a=blob_plain;hb=f62f49b4faeb477b348f51f3865dc9f4650d3029;f=index.html>Older posts</a>
  27                         </td></tr></table>
  28                         <br>
  29                         <div style="text-align:right;">
  30                         <a href="dscho.git?a=blob_plain;hb=blog;f=blog.rss"
  31                            title="Subscribe to my RSS feed"
  32                            class="rss" rel="nofollow"
  33                            style="background-color:orange;text-decoration:none;color:white;font-family:sans-serif;">RSS</a>
  34                         </div>
  35                         <br>
  36                         <table width=400px bgcolor=#e0e0e0 border=1>
  37                         <tr><th>About this blog:</th></tr>
  38                         <tr><td>
  39                         <p>It is an active <a href=http://repo.or.cz/w/git/dscho.git?a=blob_plain;f=index.html;hb=5f002cab57a837125a8f901bcd1f3c1477bc3119>abuse</a> of <a href=http://repo.or.cz/>repo.or.cz</a>,
  40                         letting gitweb unpack the objects in the current tip of the branch <i>blog</i>,
  41                         including the images and the RSS feed.
  42                         </p><p>
  43                         Publishing means running a script that collects the posts, turns them into
  44                         HTML, makes sure all the images are checked in, and pushes the result.
  45                         </p><p>
  46                         This blog also serves to grace the world with Dscho's random thoughts on and
  47                         around Git.
  48                         </p>
  49                         </td></tr></table>
  50                         <br>
  51                         <table width=400px bgcolor=#e0e0e0 border=1>
  52                         <tr><th>Links:</th></tr>
  53                         <tr><td>
  54                         <ul>
  55                         <li> <a href=http://git-scm.com/>Git's homepage</a>
  56                         <li> <a href=http://gitster.livejournal.com/>Junio's blog</a>
  57                         <li> <a href=http://www.spearce.org/>Shawn's blog</a> seems to be sitting
  58                               idle ever since he started working for Google...
  59                         <li> <a href=http://torvalds-family.blogspot.com/>Linus' blog</a> does not
  60                               talk much about Git...
  61                         <li> Scott Chacon's <a href=http://whygitisbetterthanx.com/>Why Git is better
  62                              than X</a> site
  63                         <li> <a href=http://vilain.net/>The blog of mugwump</a>
  64                         <li> <a href=http://blogs.gnome.org/newren/>Elijah Newren</a> chose the
  65                               same path as Cogito, offering an alternative porcelain (an approach
  66                               that is doomed in my opinion)
  67                         <li> <a href=http://msysgit.googlecode.com/>The msysGit project</a>, a (mostly)
  68                               failed experiment to lure the many Windows developers out there to
  69                               contribute to Open Source for a change.
  70                         </ul>
  71                         </td></tr></table>
  72                         <br>
  73                         <table width=400px bgcolor=#e0e0e0 border=1>
  74                         <tr><th>Google Ads:</th></tr>
  75                         <tr><td>
  76                         <script type="text/javascript"><!--
  77                         google_ad_client = "pub-5106407705643819";
  78                         /* 300x250, created 1/22/09 */
  79                         google_ad_slot = "6468207338";
  80                         google_ad_width = 300;
  81                         google_ad_height = 250;
  82                         //-->
  83                         </script>
  84                         <script type="text/javascript"
  85                         src="http://pagead2.googlesyndication.com/pagead/show_ads.js">
  86                         </script>
  87                         </td></tr></table>
  88                         </div>
  89                         <h6>Monday, 9th of March, Anno Domini MMIX, at the hour of the Rat</h6>
  90                         <a name=1236554268>
  91                         <h2>So, what is missing from my <i>rebase-i-p</i> branch?</h2>
  92
  93                         <p>
  94                         </p><p>
  95                         I regularly use <i>rebase -i -p</i> these days, to update my personal Git
  96                         tree (which used to be <i>my-next</i>).
  97                         </p><p>
  98                         There are a few things missing before I can start assembling a patch
  99                         series for submission:
 100                         </p><p>
 101                         <ul>
 102                         <li>I need to handle the commit parents which are outside of the rebased
 103                         ones properly.  In other words, when a commit is picked whose parent is
 104                         not rebased, it needs to be rebased onto $ONTO.
 105                         <li>The patch which uses patch-id to generate DROPPED directly also tries to
 106                         consolidate the handling of DROPPED commits by putting them into REWRITTEN
 107                         instead of DROPPED, but that breaks the tests.  So, this patch needs to be
 108                         split.
 109                         <li>I want to introduce one more command, <i>rephrase</i>, which allows you to
 110                         modify the commit message, and nothing more, and <i>halt</i>, which does the
 111                         same as <i>edit</i> without <i>pick</i>.  Then there needs to be a new test script
 112                         for those commands, and this will be an early patch series.
 113                         <li>Time.  I need time, desperately.  If my day job was not as exciting as it
 114                         is, I would have more time for Git.  As it is, I have to budget my time so
 115                         that I get anything done at all.
 116                         </ul>
 117                         </p><p>
 118                         These issues have been postponed due to Steffen taking a well-deserved
 119                         vacation, which means that I have to act as msysGit maintainer again.
 120                         </p><p>
 121                         And this coming week, I will have other things to do in all likeliness, so
 122                         that I expect to be able to submit a <i>rebase -i -p</i> patch series only next
 123                         week.  If not then, due to a heavy workload, it will be postponed to early
 124                         April.
 125                         </p><p>
 126                         Oh well, the joys of being excited by several competing projects! &#x263a;
 127                         </p>
 128                         <h6>Sunday, 8th of March, Anno Domini MMIX, at the hour of the Tiger</h6>
 129                         <a name=1236479389>
 130                         <h2>New Git for Windows version</h2>
 131
 132                         <p>
 133                         </p><p>
 134                         Phew.  That was quite a day, almost exclusively spent on finishing that
 135                         installer.  The worst part: updating GCC seemed not to be such a good idea
 136                         after all...
 137                         </p><p>
 138                         For Windows, we need to use the printf format <i>%I64u</i> (which is
 139                         non-standard, in the common way of Microsoft) if you want to print 64-bit
 140                         wide unsigned numbers.  The rest of the world accepts the standard <i>%llu</i>.
 141                         </p><p>
 142                         After upgrading to the new GCC, a lot of warnings appeared, complaining
 143                         about <i>%I64u</i>.  The warnings went away when I replaced the format with
 144                         <i>%llu</i>.
 145                         </p><p>
 146                         Being the naive I am, I mistook that for a sign that we could finally go
 147                         more standards-compliant.
 148                         </p><p>
 149                         However, it only means that we have to live with the warnings for now, as
 150                         the C runtime provided on Windows still strongly disagrees with standards
 151                         (and it has to continue to do so, lest it break existing programs).
 152                         </p><p>
 153                         Sigh.
 154                         </p><p>
 155                         At least I have the feeling that I caught the most important bugs before
 156                         releasing.
 157                         </p>
 158                         <h6>Friday, 20th of February, Anno Domini MMIX, at the hour of the Buffalo</h6>
 159                         <a name=1235092615>
 160                         <h2>Code reviews</h2>
 161
 162                         <p>
 163                         </p><p>
 164                         It has been said that reviewing patches is a most thankless job.  As I really
 165                         like the elegance of Git's source code, and care a lot about it, I did not
 166                         think that it was thankless, just a little bit tedious (especially when the
 167                         patch authors mistake criticism for personal attacks).
 168                         </p><p>
 169                         Usually, I am pretty good at ignoring insults as responses to my comments;
 170                         after all, I have a lot more enjoyable things to do than to spend time talking
 171                         to a guy who shows how wise he is when he thinks that I criticize him
 172                         <u>personally</u> when I just try to enhance his work, by offering a little bit of
 173                         my knowledge.
 174                         </p><p>
 175                         However, in the last days, three people really seemed to want to insult me,
 176                         to make me go away, to stop the fun I have with Git.
 177                         </p><p>
 178                         And they almost succeeded.
 179                         </p><p>
 180                         So I guess it is time to reassess my priorities, and maybe stop reviewing
 181                         Git patches altogether.
 182                         </p>
 183                         <h6>Thursday, 12th of February, Anno Domini MMIX, at the hour of the Tiger</h6>
 184                         <a name=1234409395>
 185                         <h2>Interactive <i>rebase</i> just learnt a new command: <i>topic</i></h2>
 186
 187                         <p>
 188                         </p><p>
 189                         Today I am pretty pleased with myself.  Two projects at my day job got a real
 190                         boost, and I implemented a shortcut that avoids the ugly 'bookmark' statement
 191                         in rebase scripts most of the time.
 192                         </p><p>
 193                         A typical rebase script, generated by <i>git rebase -i -p $COMMIT</i> will look
 194                         something like this:
 195                         </p><p>
 196                         <table
 197                                                         border=1 bgcolor=black>
 198                                                 <tr><td bgcolor=lightblue colspan=3>
 199                                 <pre>                                                                                </pre>
 200                         </td></tr>
 201                         <tr><td>
 202                                 <table cellspacing=5 border=0
 203                                          style="color:white;">
 204                                 <tr><td>
 205                                         <pre>
 206 pick 1234567 My first commit
 207 topic begin super-cool-feature
 208 pick 2345678 The super cool feature
 209 pick 3456789 Documentation for the super cool feature
 210 topic end super-cool-feature
 211 </pre>
 212                                                         </td></tr>
 213                                                         </table>
 214                                                 </td></tr>
 215                                                 </table>
 216                         </p><p>
 217                         The result will be a merge commit at the HEAD whose first parent is
 218                         "My first commit", whose second parent is "Documentation for the super
 219                         cool feature" and whose commit message is "Merge branch 'super-cool-feature'".
 220                         </p><p>
 221                         Side note: internally, <i>topic begin $NAME [at $COMMIT]</i> will be handled as if
 222                         you wrote <i>bookmark merge-parent-of-$NAME; goto $COMMIT</i>, and
 223                         <i>topic end $NAME [$MESSAGE]</i> will be handled as if you wrote
 224                         <i>bookmark $NAME; goto merge-parent-of-$NAME; merge parents $NAME [original $MARK Merge branch '$NAME']</i>.
 225                         </p><p>
 226                         Of course, being more concise, the 'topic' statement is not only nicer to the
 227                         eye, but also less error-prone.
 228                         </p><p>
 229                         And hopefully many people will agree with me that this rebase script is pretty
 230                         intuitive.
 231                         </p>
 232                         <h6>Wednesday, 11th of February, Anno Domini MMIX, at the hour of the Tiger</h6>
 233                         <a name=1234320806>
 234                         <h2>Thunderbird, oh Thunderbird, you always make my small brain hurt</h2>
 235
 236                         <p>
 237                         </p><p>
 238                         There was a lengthy discussion on the Git mailing list about using Thunderbird,
 239                         a not quite unpopular mailing program, to send inline patches.
 240                         </p><p>
 241                         It is really kind of sad that the Thunderbird developers do not see how
 242                         stubbornly they offend quite a number of people and scare them away from their
 243                         program.  After all, you should try to be liberal in what you accept and strict
 244                         in what you emit.  No, that does not mean that you should force others to
 245                         switch their mailers because you strictly adher to your philosophy in what you
 246                         emit, ignoring the rest of the world.
 247                         </p><p>
 248                         In any case, I am not affected (as long as I do not get mails from a poor soul
 249                         stuck with Thunderbird).
 250                         </p><p>
 251                         But I was a bit mean to that Thunderbird guy I dragged into the discussion, and
 252                         he seems really offended.
 253                         </p><p>
 254                         So I thought I'd give him a real reason to feel offended: I'll just do his work:
 255                         </p><p>
 256                         http://<a href=http://repo.or.cz>repo.or.cz</a>/w/UnFlowedThunderbird.git
 257                         </p><p>
 258                         It took my free time of two days, being not a Thunderbird developer myself.
 259                         Hopefully it works, and hopefully some people will feel really ashamed now.
 260                         </p>
 261                         <h6>Monday, 9th of February, Anno Domini MMIX, at the hour of the Buffalo</h6>
 262                         <a name=1234141489>
 263                         <h2><i>format-patch --thread</i> and Alpine</h2>
 264
 265                         <p>
 266                         </p><p>
 267                         I started recently to pipe the output of
 268                         <i>git format-patch --cover-letter --stdout ...</i> directly into the
 269                         <i>postponed-msgs</i> folder Alpine uses, instead of pasting files into the
 270                         mailer.
 271                         </p><p>
 272                         The idea is to pretend that I continue a postponed mail, but in reality I
 273                         never wrote it, <i>format-patch</i> did.
 274                         </p><p>
 275                         However, I had problems with the <i>--thread</i> option that is implied by
 276                         <i>--cover-letter</i>.  Alpine always generated new message IDs without adjusting
 277                         the <i>In-reply-to:</i> and <i>References:</i> headers of the other mails.
 278                         </p><p>
 279                         Now I found out that the reason is that the <i>Fcc:</i> headers were missing in
 280                         the mails, and Alpine generated them, making up new message IDs in the process.
 281                         </p><p>
 282                         Therefore I have an alias now which sets not only the <i>Fcc:</i> header, but also
 283                         the <i>To:</i> headers by rewriting the stream using <i>sed</i>.  This is slightly
 284                         ugly, but so is the handling of headers in <i>format-patch</i>: if you thought
 285                         you could specify arbitrary headers using the command line, you are mistaken:
 286                         you can do that only by editing the config.
 287                         </p><p>
 288                         While at it, I also noticed a bug whereby <i>--thread --in-reply-to=...</i> simply
 289                         forgets the <i>--thread</i>.  Maybe this week I will find time to address this bug.
 290                         </p>
 291                         <h6>Monday, 9th of February, Anno Domini MMIX, at the hour of the Buffalo</h6>
 292                         <a name=1234140696>
 293                         <h2><i>rebase</i> updates</h2>
 294
 295                         <p>
 296                         </p><p>
 297                         Phew.  The last few days, I was mainly chasing bugs I introduced due to being
 298                         too tired to work on the merge-preserving, interactive <i>rebase</i>.
 299                         </p><p>
 300                         But finally I have something I can start working with.  After my failed
 301                         experiment to use git-blame to split topic branches, I will sort the commits
 302                         in my <i>my-next</i> branch into topic branches manually.
 303                         </p><p>
 304                         Then I will add an option to <i>rebase -i -p</i> to rewrite refs which point to
 305                         rewritten commits, so that I can have branches <i>rebase-i-p</i>, <i>add-e</i>, etc
 306                         and all of them are automatically updated when I <i>rebase -i -p</i> the <i>my-next</i>
 307                         branch.
 308                         </p><p>
 309                         In the process, not only have I learnt the value of the <i>bookmark</i> command,
 310                         but made quite a few-much needed cleanups (which make
 311                         <i>git-rebase--interactive.sh</i> longer, but much more understandable).
 312                         </p><p>
 313                         Hopefully Stephan will pick the changes in the "rebase protocol" up, and then
 314                         we can have a sequencer with which I can start to make a graphical interactive
 315                         rebase using git-gui.  Or gitk.
 316                         </p><p>
 317                         Maybe.
 318                         </p>
 319                         <h6>Saturday, 7th of February, Anno Domini MMIX, at the hour of the Pig</h6>
 320                         <a name=1234040744>
 321                         <h2>The infamous <i>mark</i> command in the <i>rebase</i> command</h2>
 322
 323                         <p>
 324                         </p><p>
 325                         I realized today how easy it is to lose commits with the "merge preserving"
 326                         mode of the interactive rebase.  In my case, it was when I tried to move a
 327                         bunch of commits from the tip of my branch into a topic branch.
 328                         </p><p>
 329                         But after moving the commits, I forgot to update the parent of the merge
 330                         commit.  Possibly a mark command could have helped.  The very same command
 331                         I called a nightmare for usability.
 332                         </p><p>
 333                         So I was wrong.  Big news.  &#x263a;
 334                         </p><p>
 335                         However, I think that the syntax "mark :1" is something best left for
 336                         machine consumption, not for human beings.
 337                         </p><p>
 338                         But I have an idea: we could use some garbled commit subject, or in case of
 339                         merge parents, the merge subject as some human readable title of the mark.
 340                         </p><p>
 341                         The rebase script would then look something like this:
 342                         </p><p>
 343                         <table
 344                                                         border=1 bgcolor=black>
 345                                                 <tr><td bgcolor=lightblue colspan=3>
 346                                 <pre>                                                                                </pre>
 347                         </td></tr>
 348                         <tr><td>
 349                                 <table cellspacing=5 border=0
 350                                          style="color:white;">
 351                                 <tr><td>
 352                                         <pre>
 353 pick abcdefg Some ultra cool commit
 354 bookmark ultra-cool
 355 goto upstream
 356 pick hijklmn Some other cool commit
 357 merge parent ultra-cool Merge 'ultra-cool' into master
 358 </pre>
 359                                                         </td></tr>
 360                                                         </table>
 361                                                 </td></tr>
 362                                                 </table>
 363                         </p><p>
 364                         The good news is: I added code that refuses to finish a rebase when there
 365                         are commits that were rewritten, but not part of the new HEAD's ancestry.
 366                         </p>
 367                         <h6>Wednesday, 4th of February, Anno Domini MMIX, at the hour of the Buffalo</h6>
 368                         <a name=1233707628>
 369                         <h2>New valgrind series</h2>
 370
 371                         <p>
 372                         </p><p>
 373                         I spent quite some time cleaning up that patch series, and feel pretty
 374                         exhausted.
 375                         </p><p>
 376                         Granted, the new <i>git rebase -i -p</i> does its job without complaint so far
 377                         (so much so that I think I'll release a version of my <i>rebase</i> series
 378                         soonish), but it <u>is</u> a hassle when you have patches that you have a hard
 379                         time to decide upon the order/commit boundaries.
 380                         </p><p>
 381                         For example, I could imagine that the patch making the location of the
 382                         templates independent of the location of the Git binaries should come
 383                         <u>before</u> my patch series, and the valgrind specific part should then
 384                         be squashed into the first valgrind commit.
 385                         </p><p>
 386                         Also, it uses two features of valgrind 3.4.0:
 387                         </p><p>
 388                         <ul>
 389                         <li><i>...</i> in the suppression file, and
 390                         <li><i>--track-origins=yes</i>
 391                         </ul>
 392                         </p><p>
 393                         The latter is actually the reason I am pretty willing to keep the
 394                         requirement of that valgrind version, as it is really, really useful.
 395                         </p><p>
 396                         I guess we will see what happens to it.
 397                         </p>
 398                         <h6>Wednesday, 4th of February, Anno Domini MMIX, at the hour of the Buffalo</h6>
 399                         <a name=1233706294>
 400                         <h2>Problems with split-topic-branches.sh</h2>
 401
 402                         <p>
 403                         </p><p>
 404                         So my little script that should help me to split my topic branches does
 405                         not work properly.
 406                         </p><p>
 407                         First some background: the idea was to let <i>git blame</i> do the hard work
 408                         to find overlapping changes, i.e. changes that would conflict when
 409                         changing the order (or skipping the first change, on which the next builds).
 410                         </p><p>
 411                         The first problem with that approach: when lines are <u>removed</u> by one
 412                         commit, and the next commit touches the same location, <i>git blame</i> does
 413                         not find that the first commit is required by the second.
 414                         </p><p>
 415                         Therefore I introduced a really slow reverse thing which tries to find
 416                         those commits whose removals survived until the parent of a particular
 417                         commit, but not further.
 418                         </p><p>
 419                         However, it does not work properly.  Basically, only context sizes that
 420                         span the whole files lead to conflict-free topic branches so far.
 421                         </p><p>
 422                         As a consequence, I think I'll add an option --sprout to the revision
 423                         walker which will fake octopus merges (or a series of two-parent merges)
 424                         whenever it finds a perl of non-merge commits that are theoretically
 425                         independent, i.e. whose patches apply cleanly.
 426                         </p>
 427                 </div>
 428         </body>
 429 </html>