| Age | Commit message (Collapse) | Author | Lines |
|---|
|
Until now the shell scripts in the software building phase were in the
`reproduce/software/bash' directory. But given our recent change to a
POSIX-only start, the `configure.sh' shell script (which is the main
component of this directory) is no longer written with Bash.
With this commit, to fix that problem, that directory's name has been
changed to `reproduce/software/shell'.
|
|
Until now, the initial project scripts were primarily tested with GNU
Bash. But Bash is not generally available on all systems (it has many
features beyond POSIX). Because of this, effectively we were imposing the
requirement on the user that they must have Bash installed. We recently
started this with setting the shebang of `project' and
`reproduce/software/bash/configure.sh' to `/bin/sh'. After doing so, Raul
and Gaspar reported an error on their systems.
To fix the problem, I installed Dash (a minimalist POSIX-compliant shell)
on my computer and temporarily set the shebangs to `/bin/dash', ran the
project configuration step and fixed all issues that came up. With this
commit, it can go all the way to building GCC on my system's Dash. After
this stage (when `high-level.mk' is called), there is no problem, because
we have our own version of GNU Bash and that installed version is used.
Probably some more issues still remain and will hopefully be found in the
future.
While doing this, I also noticed the following two minor issues:
- The `./project configure' option `--input-dir' was not recognized
because it was mistakenly checking `--inputdir'. It has been corrected.
- The test C programs now use the `<<EOF' method instead of `echo'.
- In `basic.mk', the extra space between `syspath' and `:=' was removed
(it was an ancient relic!).
|
|
Until now, the hashbang of these two shell scripts was set to `/bin/bash',
hence assuming that GNU Bash exists on the host system! But this is an
extra requirement on the host operating system and these two scripts should
be written such that they operate on a POSIX shell (the generic `/bin/sh'
which can point to any shell program).
With this commit this has been implemented! We may confront some errors as
the system is run on other systems, but we should fix such errors and work
hard to make these two scripts as POSIX-compatible as possible (runnable on
any shell, so as not to force users to install Bash before running the
project).
This completes Task #15525.
|
|
Until now, the main commands to run the project were these: `./project
configure' (to build the software), `./project prepare' (to possibly
arrange input datasets and build special configuration Makefiles) and
finally `./project make' to run the project.
The main logic behind the "prepare" phase `top-prepare.mk' is to build
configuration files that can be fed into the "make" step and optimize its
operation. For example when the total number of necessary inputs for the
majority of the analysis is not as large as the total number of
inputs. With "prepare" (when necessary), you go through the raw inputs,
select the ones that are necessary for the rest of the project. The output
of `top-prepare.mk' is a configuration file (a Make variable) that keeps
the IDs (numbers, names, etc). That configuration file would then be used
in the `top-make.mk' to identify the lower level targets and allow optimal
project organization and management.
But the last two are both part of the analysis, and while they indeed need
different calls to Make to be executed, many projects don't actually need a
preparation phase: ultimately, its an implementation choice by the project
developers and doesn't concern the project users (or the developers when
they are running it).
To avoid confusing the users, or simply annoying them when a projet doesn't
need it, with this commit, the top-level `top-prepare.mk' and `top-make.mk'
Makefiles are called with the single `./project make' command and
`./project prepare' has been dropped. I noticed this while writing the
paper on this system.
|
|
Until now, it was necessry to run a long `while true' loop to see what is
currently being built at configure time. So with this commit, a new
`--checkconfig' option has been added to `./project' that can be called to
run that loop and make it easier to check.
|
|
Now that its 2020, its necessary to include this year in the copyright
statements.
|
|
Until now, after removing all environment variables, we were just giving
Make the top Makefile to execute. By default, for every target, Make will
try many built-in rules (which is good when compiling programs, but
redundant in other cases). All these checkings also populate the debugging
output of Make (with `-d'). So its easier and slightly faster to just tell
Make to ignore builtin rules and variables.
With this commit, to address this issue, the `project' script runs
`.local/bin/make' with `--no-builtin-rules' and `--no-builtin-variables'.
|
|
A special directory is now defined in `initialize.mk' that can be used in
both the preparation and build phases. Also, the contents of prepared
results can now be conditionally read during `./project make'.
|
|
In many real-world scenarios, `./project make' can really benefit from
having some basic information about the data before being run. For example
when quering a server. If we know how many datasets were downloaded and
their general properties, it can greatly optmize the process when we are
designing the solution to be run in `./project make'.
Therefore with this commit, a new phase has been added to the template's
design: `./project prepare'. In the raw template this is empty, because the
simple analysis done in the template doesn't warrant it. But everything is
ready for projects using the template to add preparation phases prior to
the analysis.
|
|
Until now, when the project's source was downloaded from something like
arXiv, in `README.md', we were instructing them to set the executable flags
of all the files that need it. But except for `./project', the reader
shouldn't have to worry about the project internals! Once its executable,
`./project' can easily fix the executable flags of all the files that need
it automatically.
With this commit, in `README.md', we just instruct the reader to set the
executable flag of `./project' and any other file that needs an executable
flag is given one at the start of the set of commands for `./project
configure'. In customized projects, if an author needs executable flags on
any other files, they can easily add it there without involving the user.
|
|
Until now, we were assuming that the users would just clone the project in
Git. But after submitting arXiv:1909.11230, and trying to build directly
from the arXiv source, I noticed several problems that wouldn't allow users
to build it automatically. So I tried the build step by step and was able
to find a fix for the several issues that came up.
The scripting parts of the fix were primarily related to the fact that the
unpacked arXiv tarball isn't under version control, so some checks had to
be put there. Also, we wanted to make it easy to remove the extra files, so
an extra `--clean-texdit' option was added to `./project'.
Finally, some manual corrections were necessary (prior to running
`./project', which are now described in `README.md'. Most of the later
steps can be automated and we should do it later, I just don't have enough
time now.
|
|
`./project make dist' will package all the LaTeX-specific files (and
analysis source files) into one `tar.gz' file that is ready to upload to
servers like arXiv. However, it wasn't updated for some time, so running it
would complain about not having a `configure' script in the top of the
project.
With this commit, it now works with the new file-structure of the project
and also copies all the BibLaTeX source files and `paper.bbl' into the top
tarball directory, which allows arXiv to build the paper as intended.
The output of `./project make dist' has been uploaded and tested on arXiv
and it is built by arXiv perfectly.
Also, a short description of all the special `make' targets was added to
the output of `./project --help'.
|
|
We recently moved the system's `rm' program absolute address to a shell
variable that is found during the `./project' script. But I had forgot to
account for the difference between the Make and Bash variable naming
differences. I had also forgot to add a value to the HOME variable.
With this commit both are corrected: the system's `rm' path is now called
`sys_rm' and the HOME variable is set.
|
|
Until now, to work on a project, it was necessary to `./configure' it and
build the software. Then we had to run `.local/bin/make' to run the project
and do the analysis every time. If the project was a shared project between
many users on a large server, it was necessary to call the `./for-group'
script.
This way of managing the project had a major problem: since the user
directly called the lower-level `./configure' or `.local/bin/make' it was
not possible to provide high-level control (for example limiting the
environment variables). This was especially noticed recently with a bug
that was related to environment variables (bug #56682).
With this commit, this problem is solved using a single script called
`project' in the top directory. To configure and build the project, users
can now run these commands:
$ ./project configure
$ ./project make
To work on the project with other users in a group these commands can be
used:
$ ./project configure --group=GROUPNAME
$ ./project make --group=GROUPNAME
The old options to both configure and make the project are still valid. Run
`./project --help' to see a list. For example:
$ ./project configure -e --host-cc
$ ./project make -j8
The old `configure' script has been moved to
`reproduce/software/bash/configure.sh' and is called by the new `./project'
script. The `./project' script now just manages the options, then passes
control to the `configure.sh' script. For the "make" step, it also reads
the options, then calls Make. So in the lower-level nothing has
changed. Only the `./project' script is now the single/direct user
interface of the project.
On a parallel note: as part of bug #56682, we also found out that on some
macOS systems, the `DYLD_LIBRARY_PATH' environment variable has to be set
to blank. This is no problem because RPATH is automatically set in macOS
and the executables and libraries contain the absolute address of the
libraries they should link with. But having `DYLD_LIBRARY_PATH' can
conflict with some low-level system libraries and cause very hard to debug
linking errors (like that reported in the bug report).
This fixes bug #56682.
|
