[PATCH 03/12] meson.build: only set build variables for non-default values

[PATCH 03/12] meson.build: only set build variables for non-default values

[Ramsay Jones]

Some preprocessor -Defines have defaults sets in the source code when
they have not been provided to the C compiler. In this case, there is
no need to pass them on the command-line, unless the build requires a
non-standard value.

The build variables for DEFAULT_EDITOR, DEFAULT_HELP_FORMAT along with
DEFAULT_PAGER have appropriate defaults ('vi', 'man' and 'less') set in
the code. Add the preprocessor -Defines to the 'libgit_c_args' only if
the values set with the corresponding 'options' are different to these
standard values.

Signed-off-by: Ramsay Jones <ramsay@ramsayjones.plus.com>
---
 meson.build | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)

diff --git a/meson.build b/meson.build
index 88a29fd043..24a4c2e3c2 100644
--- a/meson.build
+++ b/meson.build
@@ -693,10 +693,7 @@ endif
 # These variables are used for building libgit.a.
 libgit_c_args = [
   '-DBINDIR="' + get_option('bindir') + '"',
-  '-DDEFAULT_EDITOR="' + get_option('default_editor') + '"',
   '-DDEFAULT_GIT_TEMPLATE_DIR="' + get_option('datadir') / 'git-core/templates' + '"',
-  '-DDEFAULT_HELP_FORMAT="' + get_option('default_help_format') + '"',
-  '-DDEFAULT_PAGER="' + get_option('default_pager') + '"',
   '-DETC_GITATTRIBUTES="' + get_option('gitattributes') + '"',
   '-DETC_GITCONFIG="' + get_option('gitconfig') + '"',
   '-DFALLBACK_RUNTIME_PREFIX="' + get_option('prefix') + '"',
@@ -708,6 +705,16 @@ libgit_c_args = [
   '-DPAGER_ENV="' + get_option('pager_environment') + '"',
   '-DSHELL_PATH="' + fs.as_posix(shell.full_path()) + '"',
 ]
+if get_option('default_editor') != 'vi'
+  libgit_c_args += '-DDEFAULT_EDITOR="' + get_option('default_editor') + '"'
+endif
+if get_option('default_pager') != 'less'
+  libgit_c_args += '-DDEFAULT_PAGER="' + get_option('default_pager') + '"'
+endif
+if get_option('default_help_format') != 'man'
+  libgit_c_args += '-DDEFAULT_HELP_FORMAT="' + get_option('default_help_format') + '"'
+endif
+
 libgit_include_directories = [ '.' ]
 libgit_dependencies = [ ]
 
-- 
2.49.0

Re: [PATCH 03/12] meson.build: only set build variables for non-default values

[Patrick Steinhardt]

On Sat, Mar 15, 2025 at 02:46:59AM +0000, Ramsay Jones wrote:
> 
> Some preprocessor -Defines have defaults sets in the source code when
> they have not been provided to the C compiler. In this case, there is
> no need to pass them on the command-line, unless the build requires a
> non-standard value.
> 
> The build variables for DEFAULT_EDITOR, DEFAULT_HELP_FORMAT along with
> DEFAULT_PAGER have appropriate defaults ('vi', 'man' and 'less') set in
> the code. Add the preprocessor -Defines to the 'libgit_c_args' only if
> the values set with the corresponding 'options' are different to these
> standard values.

Hm. Does this really change anything though? The behaviour before and
after this patch are exactly the same as far as I understand, and by
explicitly handling the defaults we basically have to hard-code more
assumptions. So in the current form I don't see that this patch adds
much.

What we _could_ be doing is to completely drop the default values in
"meson_options.txt". In that case we could instead compare whether we
saw the empty string, which allows us to stop encoding the default vaule
both in "meson_options.txt" and in "editor.c".

Patrick

Re: [PATCH 03/12] meson.build: only set build variables for non-default values

[Ramsay Jones]

On 19/03/2025 13:36, Patrick Steinhardt wrote:
> On Sat, Mar 15, 2025 at 02:46:59AM +0000, Ramsay Jones wrote:
>>
>> Some preprocessor -Defines have defaults sets in the source code when
>> they have not been provided to the C compiler. In this case, there is
>> no need to pass them on the command-line, unless the build requires a
>> non-standard value.
>>
>> The build variables for DEFAULT_EDITOR, DEFAULT_HELP_FORMAT along with
>> DEFAULT_PAGER have appropriate defaults ('vi', 'man' and 'less') set in
>> the code. Add the preprocessor -Defines to the 'libgit_c_args' only if
>> the values set with the corresponding 'options' are different to these
>> standard values.
> 
> Hm. Does this really change anything though? The behaviour before and
> after this patch are exactly the same as far as I understand, and by
> explicitly handling the defaults we basically have to hard-code more
> assumptions. So in the current form I don't see that this patch adds
> much.

Hmm, I suppose it kinda depends on how you view it! :)

I have been looking at how the three build systems (well, mainly make
and meson) differ in various ways, in order to try and determine if
there are any significant differences and (most important) bugs.
Reducing the differences allows me to more clearly identify the bugs. ;)

In this case, the original author(s) had clearly intended that the
default values were included in the code, with the ability to override
the values from the command-line/environment only for 'non-standard' or
platform-specific uses. For example, on Windows and MINGW the
DEFAULT_HELP_FORMAT is html, which is specified in the 'config.mak.uname'
file. (I don't see this override in the meson build).

Also, the documentation (see git-var.adoc) has a statement of the compiled
in choice for the default pager and default editor, *only* if they are *not*
the standard values. I have a note, from several months ago, that says the
meson build does not pass the 'git-default-pager' and 'git-default-editor'
attributes to asciidoc. The make build only sets those attributes if the
DEFAULT_PAGER and DEFAULT_EDITOR variables are *defined* (and they should
*not* be defined to the 'standard' values or the docs would not read well).
(see git-var.adoc lines 49-51 and 67-69, Documentation/Makefile lines 239-242
and 244-247).

Also, I believe (ie I need to check) that the make build relies on the main
Makefile export-ing DEFAULT_EDITOR and DEFAULT_PAGER (see line #2923) to
make that work.

I haven't looked into all of that yet (it's one of the part #2 un-written
patches), and I don't yet know how those values get 'transmitted' to the
docs meson.build file.

Also, although I have found some meson documentation (https://mesonbuild.com/),
I haven't had the time to actually study it, so I have just used search to
try and get some answers (it seems my search-fu leaves a lot to be desired).
I was half expecting you to say something like 'hey, you don't do it like
that ... do this instead ...'. ;)

[I tried searching for a option 'is_defined()' or 'is_default()' method or
similar, but didn't find anything].

So, yes I think this patch has value. :)

ATB,
Ramsay Jones

Re: [PATCH 03/12] meson.build: only set build variables for non-default values

[Patrick Steinhardt]

On Thu, Mar 20, 2025 at 02:22:22AM +0000, Ramsay Jones wrote:
> 
> 
> On 19/03/2025 13:36, Patrick Steinhardt wrote:
> > On Sat, Mar 15, 2025 at 02:46:59AM +0000, Ramsay Jones wrote:
> >>
> >> Some preprocessor -Defines have defaults sets in the source code when
> >> they have not been provided to the C compiler. In this case, there is
> >> no need to pass them on the command-line, unless the build requires a
> >> non-standard value.
> >>
> >> The build variables for DEFAULT_EDITOR, DEFAULT_HELP_FORMAT along with
> >> DEFAULT_PAGER have appropriate defaults ('vi', 'man' and 'less') set in
> >> the code. Add the preprocessor -Defines to the 'libgit_c_args' only if
> >> the values set with the corresponding 'options' are different to these
> >> standard values.
> > 
> > Hm. Does this really change anything though? The behaviour before and
> > after this patch are exactly the same as far as I understand, and by
> > explicitly handling the defaults we basically have to hard-code more
> > assumptions. So in the current form I don't see that this patch adds
> > much.
> 
> Hmm, I suppose it kinda depends on how you view it! :)
> 
> I have been looking at how the three build systems (well, mainly make
> and meson) differ in various ways, in order to try and determine if
> there are any significant differences and (most important) bugs.
> Reducing the differences allows me to more clearly identify the bugs. ;)

Fair.

> In this case, the original author(s) had clearly intended that the
> default values were included in the code, with the ability to override
> the values from the command-line/environment only for 'non-standard' or
> platform-specific uses. For example, on Windows and MINGW the
> DEFAULT_HELP_FORMAT is html, which is specified in the 'config.mak.uname'
> file. (I don't see this override in the meson build).
> 
> Also, the documentation (see git-var.adoc) has a statement of the compiled
> in choice for the default pager and default editor, *only* if they are *not*
> the standard values. I have a note, from several months ago, that says the
> meson build does not pass the 'git-default-pager' and 'git-default-editor'
> attributes to asciidoc. The make build only sets those attributes if the
> DEFAULT_PAGER and DEFAULT_EDITOR variables are *defined* (and they should
> *not* be defined to the 'standard' values or the docs would not read well).
> (see git-var.adoc lines 49-51 and 67-69, Documentation/Makefile lines 239-242
> and 244-247).
> 
> Also, I believe (ie I need to check) that the make build relies on the main
> Makefile export-ing DEFAULT_EDITOR and DEFAULT_PAGER (see line #2923) to
> make that work.
> 
> I haven't looked into all of that yet (it's one of the part #2 un-written
> patches), and I don't yet know how those values get 'transmitted' to the
> docs meson.build file.

Okay. I very much appreciate the work that you're investing into this
area!

> Also, although I have found some meson documentation (https://mesonbuild.com/),
> I haven't had the time to actually study it, so I have just used search to
> try and get some answers (it seems my search-fu leaves a lot to be desired).
> I was half expecting you to say something like 'hey, you don't do it like
> that ... do this instead ...'. ;)
> 
> [I tried searching for a option 'is_defined()' or 'is_default()' method or
> similar, but didn't find anything].

No, there isn't anything like that. What I'd do is to set the default
values to the empty string, which makes it easy enough to see whether
the user has overridden the value to something sensible. And instead of
having the default values defined in Meson, as well, we'd be able to use
the default as specified in code.

For the `default_help_format` variable it's a bit different as it's a
combo option. But what you can do is to e.g. add a third choice
'platform' and make that the default value. We could then check for it
and either set it to the empty string on non-Windows systems so that the
default defined in our code gets used. And on Windows you'd override it
to 'html'.

Alternatively we could also refactor this option so that the default
value gets defined entirely in code and then add an empty choice.

Patrick