From joe at begriffs.com  Sun Apr 25 23:08:18 2021
From: joe at begriffs.com (Joe Nelson)
Date: Sun, 25 Apr 2021 18:08:18 -0500
Subject: Code review for new library
Message-ID: <20210425230818.GA58932@begriffs.com>

Hey all, can I get a code review on my new project?
https://github.com/begriffs/libderp

Backstory: I was working on some parser examples, and for one of the
programs I needed an associative map. I got mad that there wasn't a
library for this that worked to my liking. So I took a detour to make
(yet another) C collection library, but one that's easy to build and
avoids tricky macro magic or "invasive" data structure changes.

I haven't added any documentation yet, but thought I'd check how people
like the interface first. I tried to get inspiration from these
libraries, and find commonality:

* C++ STL https://www.cplusplus.com/reference/stl/
* Ada 2012 containers https://www.adaic.org/resources/add_content/standards/12rm/html/RM-TOC.html
* Smalltalk collection protocols https://wiki.squeak.org/squeak/uploads/172/standard_v1_9-indexed.pdf
* .NET https://docs.microsoft.com/en-us/dotnet/standard/collections/#choose-a-collection
* java.util https://docs.oracle.com/javase/7/docs/api/java/util/package-frame.html

Also consulted the Standard ML basis library, the Haskell 2010 language
report, and Common Lisp HyperSpec. However those functional libraries
didn't really apply.

So far I've implemented vector (dynamic array), list (doubly linked),
and hashmap.

Ways you can help:
* Is the current interface missing functions that you'd find useful?
  Going to add a iterator for the hashmap, for one thing...
* Are the names meaningful to you? Feel free to bikeshed.
* See any bugs? Or even have a hunch of things I should test?
* Can we simplify the Makefile in a totally portable way? It's very
  repetitive.
* The algorithms aren't fancy, but do you see obvious improvements?
  For instance, my mergesort is recursive and could theoretically
  hit stack limitations. Things along those lines.
* Does the project build on your system?
* Should the function names have a common prefix?

My plan is to rewrite the tests more methodically, using code coverage
to guide me in coming up with scenarios. Then write a man page for each
function. Then have it compile a .a and .so library. Finally add
versioning and pkg-config information, and package it up for a bunch of
OSes.

Thanks for your help!

From nicholasdrozd at gmail.com  Thu Apr 29 21:55:26 2021
From: nicholasdrozd at gmail.com (Nicholas Drozd)
Date: Thu, 29 Apr 2021 16:55:26 -0500
Subject: Code review for new library
In-Reply-To: <20210425230818.GA58932@begriffs.com>
References: <20210425230818.GA58932@begriffs.com>
Message-ID: <CABAiW0ovh1yYh8TfTU40M=CxLD4eKA_NakZVcVBBqyABE_OVhA@mail.gmail.com>

This is a very shallow review, but I couldn't get the tests to run.

  make tests
  c99 -Iinclude -DNDEBUG -O1  -o build/release/test/t_vector
build/release/vector.o test/t_vector.c
  /usr/bin/ld: cannot open output file build/release/test/t_vector: No
such file or directory
  collect2: error: ld returned 1 exit status
  make: *** [Makefile:26: build/release/test/t_vector] Error 1

That's after running `make`.

On Sun, Apr 25, 2021 at 6:08 PM Joe Nelson <joe at begriffs.com> wrote:
>
> Hey all, can I get a code review on my new project?
> https://github.com/begriffs/libderp
>
> Backstory: I was working on some parser examples, and for one of the
> programs I needed an associative map. I got mad that there wasn't a
> library for this that worked to my liking. So I took a detour to make
> (yet another) C collection library, but one that's easy to build and
> avoids tricky macro magic or "invasive" data structure changes.
>
> I haven't added any documentation yet, but thought I'd check how people
> like the interface first. I tried to get inspiration from these
> libraries, and find commonality:
>
> * C++ STL https://www.cplusplus.com/reference/stl/
> * Ada 2012 containers https://www.adaic.org/resources/add_content/standards/12rm/html/RM-TOC.html
> * Smalltalk collection protocols https://wiki.squeak.org/squeak/uploads/172/standard_v1_9-indexed.pdf
> * .NET https://docs.microsoft.com/en-us/dotnet/standard/collections/#choose-a-collection
> * java.util https://docs.oracle.com/javase/7/docs/api/java/util/package-frame.html
>
> Also consulted the Standard ML basis library, the Haskell 2010 language
> report, and Common Lisp HyperSpec. However those functional libraries
> didn't really apply.
>
> So far I've implemented vector (dynamic array), list (doubly linked),
> and hashmap.
>
> Ways you can help:
> * Is the current interface missing functions that you'd find useful?
>   Going to add a iterator for the hashmap, for one thing...
> * Are the names meaningful to you? Feel free to bikeshed.
> * See any bugs? Or even have a hunch of things I should test?
> * Can we simplify the Makefile in a totally portable way? It's very
>   repetitive.
> * The algorithms aren't fancy, but do you see obvious improvements?
>   For instance, my mergesort is recursive and could theoretically
>   hit stack limitations. Things along those lines.
> * Does the project build on your system?
> * Should the function names have a common prefix?
>
> My plan is to rewrite the tests more methodically, using code coverage
> to guide me in coming up with scenarios. Then write a man page for each
> function. Then have it compile a .a and .so library. Finally add
> versioning and pkg-config information, and package it up for a bunch of
> OSes.
>
> Thanks for your help!

From joe at begriffs.com  Fri Apr 30 02:28:31 2021
From: joe at begriffs.com (Joe Nelson)
Date: Thu, 29 Apr 2021 21:28:31 -0500
Subject: Code review for new library
In-Reply-To: <CABAiW0ovh1yYh8TfTU40M=CxLD4eKA_NakZVcVBBqyABE_OVhA@mail.gmail.com>
References: <20210425230818.GA58932@begriffs.com>
 <CABAiW0ovh1yYh8TfTU40M=CxLD4eKA_NakZVcVBBqyABE_OVhA@mail.gmail.com>
Message-ID: <20210430022831.GC67903@begriffs.com>

> This is a very shallow review, but I couldn't get the tests to run.
> 
>   make tests

Thanks for trying it out. Looks like I need better documentation. The
tests use assert() and the release build sets NDEBUG which disables
assertion checking. Try the dev variant instead:

make VARIANT=dev tests

./build/dev/test/run


(I'm curious if that'll compile for you, since the dev variant uses
clang specifically, while the release variant uses whatever your system
has assigned to CC.)

From martin.vanwinkle at gmail.com  Fri Apr 30 21:22:16 2021
From: martin.vanwinkle at gmail.com (Martin VanWinkle)
Date: Fri, 30 Apr 2021 17:22:16 -0400
Subject: Code review for new library
In-Reply-To: <20210425230818.GA58932@begriffs.com>
References: <20210425230818.GA58932@begriffs.com>
Message-ID: <CAPfkrBoT4nXVq5GKG+EKU4tEx6jqbUqZafNhi0J0P3hAn71d7g@mail.gmail.com>

Joe,

Here are the templates that I use for autotools style projects.  Man and
info pages are under doc, pkgconfig stuff is under lib.  Nothing has been
reviewed by somebody who has more experience than me, so I'm not sure if
I'm doing anything wrong.  For example, I'm pretty sure I should be using
the TESTS thing that autotools has, instead of always compiling the program
to test the library.

https://github.com/mvanwinkleias/mv_c_package_template_test/tree/master/src/templates/c_library_template2

Instructions for building debian packages are in ExtraMakefile.  I don't
have spec file generation done (yet).

The templates can be filled out by this project:

* https://github.com/theias/ias_package_shell
By specifying the project Template dir and control file like this
--project-template-dir ...
--project-control-file ...
(They're in src/templates)
(Or, you can edit the templates by hand, or I can generate them for you)

It's been a long time since I've done anything serious with C, but I'll
definitely give your code a read through.

Marty

On Sun, Apr 25, 2021, 19:08 Joe Nelson <joe at begriffs.com> wrote:

> Hey all, can I get a code review on my new project?
> https://github.com/begriffs/libderp
>
> Backstory: I was working on some parser examples, and for one of the
> programs I needed an associative map. I got mad that there wasn't a
> library for this that worked to my liking. So I took a detour to make
> (yet another) C collection library, but one that's easy to build and
> avoids tricky macro magic or "invasive" data structure changes.
>
> I haven't added any documentation yet, but thought I'd check how people
> like the interface first. I tried to get inspiration from these
> libraries, and find commonality:
>
> * C++ STL https://www.cplusplus.com/reference/stl/
> * Ada 2012 containers
> https://www.adaic.org/resources/add_content/standards/12rm/html/RM-TOC.html
> * Smalltalk collection protocols
> https://wiki.squeak.org/squeak/uploads/172/standard_v1_9-indexed.pdf
> * .NET
> https://docs.microsoft.com/en-us/dotnet/standard/collections/#choose-a-collection
> * java.util
> https://docs.oracle.com/javase/7/docs/api/java/util/package-frame.html
>
> Also consulted the Standard ML basis library, the Haskell 2010 language
> report, and Common Lisp HyperSpec. However those functional libraries
> didn't really apply.
>
> So far I've implemented vector (dynamic array), list (doubly linked),
> and hashmap.
>
> Ways you can help:
> * Is the current interface missing functions that you'd find useful?
>   Going to add a iterator for the hashmap, for one thing...
> * Are the names meaningful to you? Feel free to bikeshed.
> * See any bugs? Or even have a hunch of things I should test?
> * Can we simplify the Makefile in a totally portable way? It's very
>   repetitive.
> * The algorithms aren't fancy, but do you see obvious improvements?
>   For instance, my mergesort is recursive and could theoretically
>   hit stack limitations. Things along those lines.
> * Does the project build on your system?
> * Should the function names have a common prefix?
>
> My plan is to rewrite the tests more methodically, using code coverage
> to guide me in coming up with scenarios. Then write a man page for each
> function. Then have it compile a .a and .so library. Finally add
> versioning and pkg-config information, and package it up for a bunch of
> OSes.
>
> Thanks for your help!
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://talk.begriffs.com/pipermail/friends/attachments/20210430/e55658ce/attachment.html>