Opened 10 years ago
Closed 10 years ago
#12898 closed defect (fixed)
Update top-level README.txt
Reported by: | jdemeyer | Owned by: | mvngu |
---|---|---|---|
Priority: | blocker | Milestone: | sage-5.0 |
Component: | documentation | Keywords: | |
Cc: | drkirkby | Merged in: | sage-5.0.rc1 |
Authors: | Jeroen Demeyer | Reviewers: | William Stein, Karl-Dieter Crisman, David Kirkby, Martin Raum |
Report Upstream: | N/A | Work issues: | |
Branch: | Commit: | ||
Dependencies: | Stopgaps: |
Description
The file $SAGE_ROOT/README.txt
still says you need a fortran compiler, except on OSX, with no
mention of the gcc package, or how one might use or disable it.
Attachments (1)
Change History (21)
comment:1 Changed 10 years ago by
- Status changed from new to needs_review
comment:2 follow-up: ↓ 3 Changed 10 years ago by
- Status changed from needs_review to needs_work
This doesn't make sense.
Since includes a GCC (GNU Compiler Collection) package. In order to
Trivial to fix but I don't know exactly what wording you want here. Maybe just one sentence instead of two? Anyway, the first sentence has no verb.
comment:3 in reply to: ↑ 2 Changed 10 years ago by
comment:4 Changed 10 years ago by
- Status changed from needs_work to needs_review
comment:5 Changed 10 years ago by
- Status changed from needs_review to positive_review
It looks good to me.
comment:6 follow-up: ↓ 8 Changed 10 years ago by
Aah! I was in the process of submitting...
For reference:
Anyway, the first sentence has no verb.
"includes" is a verb :-)
Main verb, yes.
Questions/comments:
- Some things I kind of get are true from previous tickets (3 GB, changing to
--optional
from-optional
and so forth) but I don't actually know they are true; I guess I should just trust the author? ;-) - So Sage compiles with Sun compilers on Solaris now because we build and use GCC? I don't remember reading that but haven't been paying close attention to the thread, seems plausible.
- I understand that the
v
flag for untarring is optional, but if you remove that flag, then you should probably tell the user how long they should expect to wait for Sage to untar. It takes a long time (I've done this with a student on a couple occasions and we weren't sure if the connection had timed out!), and I would expect that some of the spkgs inside Sage likewise take at least more than a couple minutes to untar - ?
But in general this looks good and applies properly etc. Positive review? I can't take responsibility for every one of the little changes, like the comment about Ubuntu 12.04, obviously, so I hesitate... but seems like it's ready and certainly helpful.
comment:7 Changed 10 years ago by
- Reviewers set to William Stein, Karl-Dieter Crisman
comment:8 in reply to: ↑ 6 ; follow-up: ↓ 9 Changed 10 years ago by
Replying to kcrisman:
So Sage compiles with Sun compilers on Solaris now because we build and use GCC?
Exactly. It's even tested on the buildbot (on i386 OpenSolaris at least).
I understand that the
v
flag for untarring is optional, but if you remove that flag, then you should probably tell the user how long they should expect to wait for Sage to untar. It takes a long time (I've done this with a student on a couple occasions and we weren't sure if the connection had timed out!), and I would expect that some of the spkgs inside Sage likewise take at least more than a couple minutes to untar - ?
Fair enough, I added back that flag.
comment:9 in reply to: ↑ 8 Changed 10 years ago by
I understand that the
v
flag for untarring is optional, but if you remove that flag, then you should probably tell the user how long they should expect to wait for Sage to untar. It takes a long time (I've done this with a student on a couple occasions and we weren't sure if the connection had timed out!), and I would expect that some of the spkgs inside Sage likewise take at least more than a couple minutes to untar - ?Fair enough, I added back that flag.
Either that or a clarification without -v
is fine, I just wasn't sure if there was some other reason to get rid of that flag (e.g., the verbosity for some with a lot of image files is VERY verbose). I assume the -xj
and -jx
switch is because the mandatory flag is supposed to come first? I know little of such things but that seems plausible.
Nice that we can build more places. I wonder if Sage would even build with other compilers, then... say, on Windows or some other Unices... but probably not, due to other issues like linkers etc.? (About which I also know nothing, but it's fun to speculate.)
comment:10 follow-up: ↓ 11 Changed 10 years ago by
Is is true that "Each spkg in $SAGE_ROOT/spkg/standard/ is a bzip'd tarball", as stated? I doubt the bzip2 package is. I think that is an uncompressed tarball. I'm not sure if there are any others either.
Also, I don't know why anyone puts -xfj or similar, when xfj will do. Tar is not defined by POSIX, so it's impossible to give an authoritative answer, but every tar program I have ever used, does not need the minus sign. The use of x (to extract), c (to create), t (to view), u (to update) or r (to replace) are not options. You must have one of them. Since they are not options, you don't need the minus sign. The j is an option, but GNU tar does not require the minus sign. So one might as well save a bit of wear on the keyboard, and not put the minus sign.
I thought GNU tar was supposed to be a requirement to build Sage. In which case, one might as well drop the j too, since GNU tar will work out for itself that the file is compressed or not. If it's compressed, it will silently uncompress it.
Dave
comment:11 in reply to: ↑ 10 Changed 10 years ago by
Replying to drkirkby:
Is is true that "Each spkg in $SAGE_ROOT/spkg/standard/ is a bzip'd tarball", as stated? I doubt the bzip2 package is. I think that is an uncompressed tarball. I'm not sure if there are any others either.
bzip2
is gzip-compressed, the other packages are all bzip2-compressed.
comment:12 Changed 10 years ago by
- Status changed from positive_review to needs_work
David, I made some changes you suggested. Please review.
comment:13 Changed 10 years ago by
- Reviewers changed from William Stein, Karl-Dieter Crisman to William Stein, Karl-Dieter Crisman, David Kirkby
- Status changed from needs_work to needs_review
comment:14 Changed 10 years ago by
new version says : "recent and equal versions of gcc, g++ and gfortran."
It is not clear what "equal" means here...
comment:15 Changed 10 years ago by
Clarified:
If you don't want to install GCC, you need to have recent versions of gcc, g++ and gfortran; moreover, the versions must be equal.
comment:16 Changed 10 years ago by
I think the double space looks a bit too much like emacs, and I removed it. I also made the link to the environment variable page more visible, and I removed the dash in the second tar command, just to be a bit more consequent (the other was removed as well).
Apart from this the changes look good.
comment:17 Changed 10 years ago by
I made some further changes:
-
README.txt
diff --git a/README.txt b/README.txt
a b 69 69 70 70 1. Make sure you have the dependencies and 3 GB of free disk space. 71 71 72 Linux: GCC, make, m4, perl, ranlib, and tar.72 Linux: gcc, make, m4, perl, ranlib, and tar. 73 73 (install these using your package manager) 74 74 On recent Debian or Ubuntu systems (in particular Ubuntu 12.04 75 75 "Precise"), you need the dpkg-dev package. 76 76 77 OS X: Xcode. 77 OS X: Xcode. Make sure you have installed the most recent version 78 78 of Xcode. For pre-Lion versions of OS X, you can download Xcode 79 79 from http://developer.apple.com/downloads/. For OS X Lion, you can 80 80 install it using the App Store. With Xcode 4.3 or later, you need … … 112 112 113 113 On Linux, if you get this error message: 114 114 115 " restore segment prot after reloc: Permission denied "115 Error: cannot restore segment prot after reloc: Permission denied 116 116 117 117 the problem is probably related to SELinux. See the following URL for 118 118 further information: 119 119 120 http://www. ittvis.com/services/techtip.asp?ttid=3092120 http://www.exelisvis.com/Support/HelpArticleDetail/ArticleId/3092.aspx 121 121 122 122 123 123 IMPLEMENTATION … … 184 184 export SAGE_CHECK="yes" 185 185 186 186 before starting the Sage build. This will run each test suite and 187 will raise an error if any failures occur. Warning: on many 188 platforms, this will cause failures in the installation of the 189 Python spkg, so Python's test suite has been disabled by default. 190 To renable it, set the environment variable SAGE_CHECK_PACKAGES to 191 'python'. 192 187 will raise an error if any failures occur. Python's test suite has 188 been disabled by default, because it causes failures on most 189 systems. To renable the Python testsuite, set the environment 190 variable SAGE_CHECK_PACKAGES to "python". 191 193 192 To start the build, type: 194 193 195 194 make 196 195 197 198 196 4. Wait about 1 hour to 14 days, depending on your computer (it took 199 197 about 2 weeks to build Sage on the T-Mobile G1 Android cell phone). 200 198 … … 265 263 266 264 Sage includes a GCC (GNU Compiler Collection) package. In order to 267 265 build Sage, you need a C compiler which can build GCC and its 268 prerequisites. GCCversion 4.0.1 or later should probably work. On266 prerequisites. gcc version 4.0.1 or later should probably work. On 269 267 Solaris or OpenSolaris, building with the Sun compiler should also work. 270 268 271 269 The GCC package in Sage is not always installed. It is determined
Changed 10 years ago by
comment:18 Changed 10 years ago by
- Status changed from needs_review to positive_review
Looks good now.
comment:19 Changed 10 years ago by
- Reviewers changed from William Stein, Karl-Dieter Crisman, David Kirkby to William Stein, Karl-Dieter Crisman, David Kirkby, Martin Raum
comment:20 Changed 10 years ago by
- Merged in set to sage-5.0.rc1
- Resolution set to fixed
- Status changed from positive_review to closed
This is a first attempt.