readme links
tombstone
content sync'd with emacs-geiser/geiser 0.13
we need macros.texi
macros.texi removed
texi updates
Merge remote-tracking branch 'gitlab/master'
user manual and web pages updates
install.texi
geiser-mit.el deleted
geiser-guile moved to its own package
geiser-chicken moved to its own package
geiser-gambit moved to its own package
geiser-chibi moved to its own package
geiser-chez moved to its own package
geiser-racket moved to individual package
geiser-mit moved to individual package
geiser-messages-mode: Don't quote lambda expressions
Racket: macro-expansion: don't quote result
`pretty-print' by default quotes its output, so that `eval' can yield the
correct value. This is wrong in macro-expansions though, since the quote
wouldn't be there in the actual expanded code, and is also inconsistent with
other REPLs (e.g Guile).
Use a quote-depth of 1, which doesn't cause the extra quote to be printed for
macro expansions (`read' yields the correct value).
`pretty-print' by default quotes its output, so that `eval' can yield the
correct value. This is wrong in macro-expansions though, since the quote
wouldn't be there in the actual expanded code, and is also inconsistent with
other REPLs (e.g Guile).
Use a quote-depth of 1, which doesn't cause the extra quote to be printed for
macro expansions (`read' yields the correct value).
Determine the display string of command key dynamically.
* elisp/geiser-repl.el(geiser-repl--sentinel): Use `substitute-command-keys'.
* elisp/geiser-repl.el(geiser-repl--sentinel): Use `substitute-command-keys'.
fix problems in module detection code (chicken); remove micro-optimisation
- fix typo that renamed the regex from "module" to "define-module"
- revert previous change that removed support for single-identifier module
names
- as suggested, remove looking-at micro-optimisation, since it make code harder
to read and doesn't seem to bring any relevant benefits.
- fix typo that renamed the regex from "module" to "define-module"
- revert previous change that removed support for single-identifier module
names
- as suggested, remove looking-at micro-optimisation, since it make code harder
to read and doesn't seem to bring any relevant benefits.
Fix error when compiling define-library form in Chicken
Problem: when running geiser-compile-current-buffer (C-c C-k) on a file
using R7RS's define-library form, geiser-chicken--get-module starts an
infinite recursion. This happens because the original regular expression
contains two groups, but the code only tries to fetch the library name
from the first one with "(match-string-no-properties 1)".
Solution: we use the same strategy as in geiser-guile.el, namely we split the
regular expression into two separate ones, and check them one by one.
Problem: when running geiser-compile-current-buffer (C-c C-k) on a file
using R7RS's define-library form, geiser-chicken--get-module starts an
infinite recursion. This happens because the original regular expression
contains two groups, but the code only tries to fetch the library name
from the first one with "(match-string-no-properties 1)".
Solution: we use the same strategy as in geiser-guile.el, namely we split the
regular expression into two separate ones, and check them one by one.
Some comment and whitespace cleanup
Add doc-string to geiser-custom--defcustom
Also add a FIXME comment about how this macro isn't actually needed.
Also add a FIXME comment about how this macro isn't actually needed.
Improve doc-strings of some options
The whole first sentence should fit on the first line. If that makes
the line a bit long then that is unfortunate but better than wrapping
it onto a new line. When wrapping onto a new line anyway then the
second line should never be intended. When it can be avoided, then
long first lines should be made shorter.
The whole first sentence should fit on the first line. If that makes
the line a bit long then that is unfortunate but better than wrapping
it onto a new line. When wrapping onto a new line anyway then the
second line should never be intended. When it can be avoided, then
long first lines should be made shorter.
Declare keymaps using defvar explicitly
This makes it possible to re-evaluate the containing
buffers without user customizations being clobbered.
This makes it possible to re-evaluate the containing
buffers without user customizations being clobbered.
Placate byte-compiler
guile: Recognize library declaration correctly
Support buffer-local binary and arglist configuration
Retrieve the scheme binary and arglist before switching to the repl
buffer, so that we respect any local settings from the source buffer.
Retrieve the scheme binary and arglist before switching to the repl
buffer, so that we respect any local settings from the source buffer.
Gambit version now contains also the commit id. Code updated accordingly
External help fix
Fixes an issue where chicken 5 users were taken to the chicken 4 external docs for the given query. With this update, users will be taken to chicken 5 or 4 docs based on the result of (geiser-chicken--version).
Fixes an issue where chicken 5 users were taken to the chicken 4 external docs for the given query. With this update, users will be taken to chicken 5 or 4 docs based on the result of (geiser-chicken--version).
Version 0.12
Always display error in minibuffer after eval
When evaluating expressions in a Scheme buffer, display the error
of an evaluation in the minibuffer regardless of whether
`geiser-debug-show-debug-p` or `geiser-debug-jump-to-debug-p` are
set or not.
When evaluating expressions in a Scheme buffer, display the error
of an evaluation in the minibuffer regardless of whether
`geiser-debug-show-debug-p` or `geiser-debug-jump-to-debug-p` are
set or not.
chez: pretty printing macroexpand results.
make chez display output and error messages.
Add docs for per-project repl instances
Tag project functions instead of overriding docs
Fix byte-compiler warning
Improve project function selection, make ignore default
Refactor to make repl/impl proj argument optional
Use a separate REPL instance per project
add geiser# prefix to geiser calls
Previously after switching to a module with say ",m my-module", evaluation of
geiser commands from a buffer failed with:
Error: unbound variable: my-module#geiser-eval
A suggestion on the #chicken channel was to prefix all geiser calls with the
module name "geiser#". I side-effect of this is that we don't need
to import geiser anymore, reducing "pollution" of the user's code with geiser
symbols.
This commit adds the geiser# prefix to geiser calls, and remove "(import geiser)" when loading chicken 5. The same could be done for chicken 4, I didn't have
time to change and test it though.
Previously after switching to a module with say ",m my-module", evaluation of
geiser commands from a buffer failed with:
Error: unbound variable: my-module#geiser-eval
A suggestion on the #chicken channel was to prefix all geiser calls with the
module name "geiser#". I side-effect of this is that we don't need
to import geiser anymore, reducing "pollution" of the user's code with geiser
symbols.
This commit adds the geiser# prefix to geiser calls, and remove "(import geiser)" when loading chicken 5. The same could be done for chicken 4, I didn't have
time to change and test it though.
Returning C-c k back to users, that key's theirs!
Should close issue #315, where it is pointed out that "sequences
consisting of `C-c` and a letter (either upper or lower case) are
reserved for users; they are the *only* sequences reserved for users,
so do not block them."
Should close issue #315, where it is pointed out that "sequences
consisting of `C-c` and a letter (either upper or lower case) are
reserved for users; they are the *only* sequences reserved for users,
so do not block them."
Change geiser-eval to accept lists as module name (C5)
Chicken 5 allows lists of symbols and integers as module names, for instance
(module (my example) * ...)
see: http://wiki.call-cc.org/man/5/Modules#module
Chicken 5 allows lists of symbols and integers as module names, for instance
(module (my example) * ...)
see: http://wiki.call-cc.org/man/5/Modules#module
Links to olde blog posts
Thanks to Sebastian, who noted this in issue #314
Thanks to Sebastian, who noted this in issue #314
geiser-chicken--version: Don't use a shell
Using shell-command-to-string requires making assumptions about how
quoting works in the shell. While single quotes are okay for bourne
shells, it doesn't work for cmd.exe (the default on Windows).
Use call-process instead to remove dependency on shell quoting
details.
Using shell-command-to-string requires making assumptions about how
quoting works in the shell. While single quotes are okay for bourne
shells, it doesn't work for cmd.exe (the default on Windows).
Use call-process instead to remove dependency on shell quoting
details.
Fix typos
Make guile return line numbers more often, as suggested by Sergey
See gitlab issue #303, where Sergey Trofimov kindly described not only
the symptons, but this cure.
See gitlab issue #303, where Sergey Trofimov kindly described not only
the symptons, but this cure.
Better handcrafted reader, now with numbers!
For no good reason, we were transforming numbers in retorts to symbols
and then failing to recognise things like line or column numbers
there.
The "fix" here only works for the intersection of numbers that are
written in the same way in elisp and scheme; one day we'll find a
situation where this doesn't cut it, but right now we only really use
integers.
For no good reason, we were transforming numbers in retorts to symbols
and then failing to recognise things like line or column numbers
there.
The "fix" here only works for the intersection of numbers that are
written in the same way in elisp and scheme; one day we'll find a
situation where this doesn't cut it, but right now we only really use
integers.
Completion for vanilla users
Restoring what seem reasonable definitions for the completion
functions when called away from the current buffer. Should fix issue
Restoring what seem reasonable definitions for the completion
functions when called away from the current buffer. Should fix issue
Fix typos
add extra parameter support to chez
add extra parameter support to chez implementation
Fix recently added .dir-locals.el
Chop off trailing whitespace before printing REPL result
Use cl-lib instead of cl
Starting with Emacs 27 cl is fully deprecated, including at
compile-time.
Starting with Emacs 27 cl is fully deprecated, including at
compile-time.
Declare geiser-restart-repl in geiser-compile.el
Delete trailing whitespace
Fix indentation
Enforce use of spaces for indentation
We cannot just assume that each and every contributors
has independently configured their Editor that way.
We cannot just assume that each and every contributors
has independently configured their Editor that way.
Mark the beginning of code part of elisp libraries with Code: heading
It's the convention and by following it we make a big step towards
supporting outline navigation.
The convention doesn't say much about what parts of the code are
supposed to be part of that sections and what parts belong in a
subsequent section. Here we put the `require' forms in this section
and maybe some setup code, that's a popular approach.
In most cases there was "\f" where we now insert "Code:". They both
serve a similar purpose and we keep the former because some users
depend on that for navigation. We even add this "\f" in libraries
where it previously was missing.
In some cases the permission statement was followed by a commentary,
which obviously does not belong in the "Code:" section. In such cases
add the conventional "Commentary:" section.
It's the convention and by following it we make a big step towards
supporting outline navigation.
The convention doesn't say much about what parts of the code are
supposed to be part of that sections and what parts belong in a
subsequent section. Here we put the `require' forms in this section
and maybe some setup code, that's a popular approach.
In most cases there was "\f" where we now insert "Code:". They both
serve a similar purpose and we keep the former because some users
depend on that for navigation. We even add this "\f" in libraries
where it previously was missing.
In some cases the permission statement was followed by a commentary,
which obviously does not belong in the "Code:" section. In such cases
add the conventional "Commentary:" section.
Begin the summary lines of all elisp libraries with three semicolons
It's the convention and by following it we make a big step towards
supporting outline navigation.
It's the convention and by following it we make a big step towards
supporting outline navigation.
Remove duplicated definition of geiser-connect
define-geiser-implementation - un-unquote load-file-name
Cleaner Chicken and Gambit sections in install docs
Documentation fixes
Version bump
Fix for misspelled geiser-completion-module-list-func (#271)
Not referring to savannah downloads in the docs
What with MELPA and whatever is happening with savannah ssh access.
What with MELPA and whatever is happening with savannah ssh access.
Version bump
Whitespace
Add Gambit source files to makefiles
doc: Improve macros to add docbook support
Refix #271, and consider gardening instead of programming
Guile: use load paths defined in dir-local variables (fixes #268)
A similar idea should probably be used with other schemes, but right
now i feel ashamed of having taken so long to fix this one (assuming
it's fixed!), so let's rush this commit for a change.
A similar idea should probably be used with other schemes, but right
now i feel ashamed of having taken so long to fix this one (assuming
it's fixed!), so let's rush this commit for a change.
Proper wrapping of final comment lines (fixes #292)
Thanks a lot Sean Delvin for a great bug report which, moreover,
contained the solution to the problem! (even though i'm risking a
small modification).
Thanks a lot Sean Delvin for a great bug report which, moreover,
contained the solution to the problem! (even though i'm risking a
small modification).
Compatibility with Helm completions in emacs-25 (fixes #271)
Version 0.11
Really fix #252
And by a new pair of glasses in the process.
And by a new pair of glasses in the process.
Leftover (fixes #291)
Repeat with me: try M-x geiser-reload before pushing to gitlab
Repeat with me: try M-x geiser-reload before pushing to gitlab
Merge branch 'support-mit-scmutils' [279]
C-c \ and C-c [ in geiser-repl too
With a hat tip to Mikhail Kryshen, who was wondering in guile-user why
oh why, and rightly so.
With a hat tip to Mikhail Kryshen, who was wondering in guile-user why
oh why, and rightly so.
Add rudimentary Chez support for macro expansion
Support scmutils, a big library for mit-scheme
mit-scheme is case-insensitive by default, but scmutils is case-sensitive,
so changing argument to lowercase to support both.
mit-scheme is case-insensitive by default, but scmutils is case-sensitive,
so changing argument to lowercase to support both.
Capture stdout in ChezScheme's eval:geiser
Rename multiline expression option
* Renames geiser-repl-context-sensitive-send to
geiser-repl-send-on-return-p. This option's value is now inverted.
* Update documentation accordingly.
* Renames geiser-repl-context-sensitive-send to
geiser-repl-send-on-return-p. This option's value is now inverted.
* Update documentation accordingly.
elisp/geiser-completion.el: Expose `--symbol-list` and `--module-list` funcs.
Merge branch 'master' into ChibiExperiments
Add printing a stack trace to the user code exceptions.
Automatically enable racket for .rktd and .rktl files.
elisp/geiser.doc.el: Add geiser-doc-manual-for-symbol.
Missing require (fixes issue #285)
Seems we forgot a require while adding a new defcustom in geiser-log.
Seems we forgot a require while adding a new defcustom in geiser-log.
whitespace
Add a few minor tweaks.
Allow additional parameters to Chibi binary
Fix latest chicken. Add debug level logging. Fixed typo.
Add a custom variable to scroll *geiser messages*.
Add a (declare (debug)) form to all geiser-%s--with-buffer macros.
Set header-line-format only in case when `cond` branch is true
fix the output redirection of gambit scheme interpreter
the gsi's output wasn't redirected to the right place
the gsi's output wasn't redirected to the right place