1 |
Duncan wrote: |
2 |
> "John P. Burkett" <burkett@×××.edu> posted 4A06F178.7060007@×××.edu, |
3 |
> excerpted below, on Sun, 10 May 2009 11:23:36 -0400: |
4 |
> |
5 |
>> Duncan wrote: |
6 |
>>> "John P. Burkett" <burkett@×××.edu> posted 49FDCD09.7070204@×××.edu, |
7 |
>>> excerpted below, on Sun, 03 May 2009 12:57:45 -0400: |
8 |
>>> |
9 |
>>>> Thanks, Duncan. [...] The results are the same; the latest version |
10 |
>>>> available version of sys-apps/portage is still listed as 2.1.6.11, |
11 |
>>>> which is the version I have installed. |
12 |
>>>> |
13 |
>>>> I attempted to manually download the source file, and place it in |
14 |
>>>> distfiles, and then run emerge. [Still had the problem.] |
15 |
>>> That's clearly a portage bug (even if we didn't already know it based |
16 |
>>> on the bug you mentioned and the new versions that are /supposed/ to be |
17 |
>>> out), as that argument list isn't even that long at all. |
18 |
>>> |
19 |
>>> So one way or another, we gotta get around that bug. |
20 |
>>> |
21 |
>>> You've verified that you can decompress that source archive manually, |
22 |
>>> right? |
23 |
> |
24 |
>> Yes, it appears that I can decompress lzma files. |
25 |
> |
26 |
>>> Meanwhile, on portage upgrade side... |
27 |
> |
28 |
>>> 2.1.6.12 is indeed in-tree, but no arch has keyworded it stable yet. |
29 |
>>> I don't see any masking and checking the ebuild itself, I see it's |
30 |
>>> keyworded ~arch. |
31 |
> |
32 |
>>> I don't know why it hasn't been stable-keyworded, except that |
33 |
>>> archs probably haven't gotten to it yet, but you might wish to consider |
34 |
>>> adding: |
35 |
>>> |
36 |
>>> ~sys-apps/portage-2.1.6.12 |
37 |
>>> |
38 |
>>> ... to your package.keywords file or directory. |
39 |
> |
40 |
>> After adding ~sys-apps/portage-2.1.6.12 to my package.keywords file, I |
41 |
>> did "emerge portage". That process appears to have been successful. Now |
42 |
>> when I do "emerge --search portage" the response is |
43 |
>> * sys-apps/portage |
44 |
>> Latest version available: 2.1.6.12 |
45 |
>> Latest version installed: 2.1.6.12 |
46 |
> |
47 |
> OK, great. At least that's still working as it should. |
48 |
> |
49 |
>> So far, so good. However, when I do "emerge texlive-latexextra", the |
50 |
>> response is as follows: |
51 |
>> Calculating dependencies... done! |
52 |
>> |
53 |
>>>>> Verifying ebuild manifests |
54 |
>>>>> Emerging (1 of 1) dev-texlive/texlive-latexextra-2008-r1 |
55 |
>> [Errno 7] Argument list too long: |
56 |
> |
57 |
> [etc] |
58 |
> |
59 |
>>> There are other alternatives too. Did you try using the --fetchonly |
60 |
>>> option? The bug mentions that worked for some people. |
61 |
>> Doing "emerge -f texlive-latexextra" also produces "argument list too |
62 |
>> long" errors, for example: |
63 |
>>>>> Downloading |
64 |
>> 'http://ftp.jaist.ac.jp/pub/Linux/Gentoo/distfiles/texlive-module- |
65 |
> pdfcprot-2008.tar.lzma' |
66 |
>> [Errno 7] Argument list too long: |
67 |
>> /usr/bin/wget -t 5 -T 60 --passive-ftp -O |
68 |
>> /usr/portage/distfiles/texlive-module-pdfcprot-2008.tar.lzma |
69 |
>> http://ftp.jaist.ac.jp/pub/Linux/Gentoo/distfiles/texlive-module- |
70 |
>> pdfcprot-2008.tar.lzma |
71 |
> |
72 |
> So the workaround others found doesn't work for you. Ugly! |
73 |
> |
74 |
>>> There's some additional discussion on why it happens -- are you using |
75 |
>>> an old kernel (<2.6.23)? They had shorter max commandline lengths. |
76 |
>>> Thus, upgrading your kernel is presumably another alternative. |
77 |
> |
78 |
>> I'm using kernel 2.6.20-gentoo-r6. Upgrading to a more recent kernel |
79 |
>> would probably be beneficial but may require skills that I lack. |
80 |
> |
81 |
> Well, you'll need to learn to do it at some point. Now's probably as |
82 |
> good a time as any... |
83 |
> |
84 |
> OK. We have two paths we can work on. First, whatever they did to fix |
85 |
> the bug in portage-2.1.6.12 might have fixed it for some, but it didn't |
86 |
> for you. Thus, one path is continuing to work on the portage bug. At |
87 |
> this point, it's time to reopen (or clone since we can't reopen) the bug |
88 |
> and get the portage devs looking at it again, since it's a portage bug |
89 |
> that's obviously not fixed for you, when they think it is. |
90 |
> |
91 |
> I see 2.1.6.13 is out now. You will want to update to it (you'll need to |
92 |
> change your package.keywords again) to ensure the bug isn't fixed by |
93 |
> something else they did. If the bug is still there, here's what I'd do. |
94 |
> |
95 |
> Go to bug 262647. As it isn't your bug, you can't reopen it. However, |
96 |
> you can clone it (link at the bottom of the main bug status area, under |
97 |
> the Additional Comments textbox, but BEFORE the comments, "Clone This |
98 |
> Bug" link to the right). Do so, choose as a product "Portage |
99 |
> Development", version 2.1, component "Core", then explain why you are |
100 |
> cloning the bug, that 2.1.6.12 (and .13) that were supposed to have the |
101 |
> fix, don't seem to work for you. |
102 |
> |
103 |
> You'll also want to mention that -f and/or downloading it manually to |
104 |
> $DISTFILES, then rerunning the emerge, does NOT fix it for you. Mention |
105 |
> your kernel version 2.6.20 as that's useful information as well, add your |
106 |
> emerge --info and anything else you can think of that would be useful. |
107 |
> |
108 |
> You can also link to this thread as it's seen on gmane, so they can see |
109 |
> what we've already tried without you having to retype it all. Finally, |
110 |
> add me to the CC for the bug, as I'm interested in following it. |
111 |
> |
112 |
> http://comments.gmane.org/gmane.linux.gentoo.amd64/14380 |
113 |
|
114 |
Thank you, Duncan, for your very clear and helpful suggestions. |
115 |
I have now cloned the bug. It is in the database as bug 270230. The CC |
116 |
list includes 1i5t5.duncan@×××.net. |
117 |
|
118 |
While waiting for a portage bug fix, I'll follow your suggestions about |
119 |
trying to learn to update my kernel. Thanks again! |
120 |
|
121 |
|
122 |
|
123 |
> |
124 |
> Meanwhile, the second path we can try is the kernel upgrade path. You |
125 |
> weren't sure about it, but luckily, Gentoo has quite some help on the |
126 |
> subject. But first, a hint. Gentoo really does have some good |
127 |
> documentation. Actually, Gentoo's documentation is often good enough |
128 |
> folks from other distributions use it too. =:^) As such, anything big |
129 |
> you're thinking about doing on your system, it's worth checking to see |
130 |
> Gentoo has some documentation on the subject. What I usually do is go to |
131 |
> the docs list page and then use my browser's search function look for |
132 |
> docs on whatever it is I'm interested in. In this case, that's kernel |
133 |
> upgrades, so I searched on "kernel", and came up with the below list of |
134 |
> links from the documents list page. |
135 |
> |
136 |
> The big list of Gentoo documents. Bookmark it! =:^) |
137 |
> |
138 |
> http://www.gentoo.org/doc/en/list.xml |
139 |
> |
140 |
> ..... |
141 |
> |
142 |
> Some kernel related documentation that should be of help: |
143 |
> |
144 |
> Gentoo Linux Kernel Upgrade Guide |
145 |
> |
146 |
> The kernel upgrade guide looks to be the one most immediately of interest |
147 |
> here. |
148 |
> |
149 |
> http://www.gentoo.org/doc/en/kernel-upgrade.xml |
150 |
> |
151 |
> |
152 |
> Gentoo Linux Kernel Guide |
153 |
> |
154 |
> The kernel guide is mainly about choosing the right kernel sources, from |
155 |
> multiple choices available. |
156 |
> |
157 |
> http://www.gentoo.org/doc/en/gentoo-kernel.xml |
158 |
> |
159 |
> |
160 |
> Gentoo Linux Kernel Configuration Guide |
161 |
> |
162 |
> The kernel config guide is an overview of the manual kernel configuration |
163 |
> process. It doesn't go into too much detail, but if you've never done it |
164 |
> before, it could be very helpful, none-the-less. It does cover several |
165 |
> areas people often find confusing, SATA as SCSI, IDE and DMA, USB, multi- |
166 |
> core/multi-CPU, etc. Another section links a bunch of other more topic |
167 |
> specific documents, on ALSA/sound, bluetooth, printing, power management, |
168 |
> USB, etc. |
169 |
> |
170 |
> http://www.gentoo.org/doc/en/kernel-config.xml |
171 |
> |
172 |
> |
173 |
> Compiling the Linux kernel |
174 |
> |
175 |
> This is not a Gentoo-specific document but one written originally for IBM |
176 |
> developerWorks by Daniel Robbins, Gentoo's founder. It'll give you a lot |
177 |
> of generally useful background and DRobbins is good at explaining things, |
178 |
> but parts of it are now a bit dated. For example, it discusses LILO for |
179 |
> booting, while Gentoo (and most modern x86 and amd64 distributions) now |
180 |
> uses GRUB. Thus, you'll probably want to skip that section (section 5). |
181 |
> |
182 |
> http://www.gentoo.org/doc/en/articles/linux-kernel-compiling.xml |
183 |
> |
184 |
> |
185 |
> The Gentoo Handbook also covers the kernel briefly, in the install |
186 |
> coverage. But you probably read that while installing and are still a |
187 |
> bit leery about kernel config, thus the above. I'm mentioning it here |
188 |
> mainly for completeness, but if you've not looked at it in awhile, it |
189 |
> couldn't hurt to review that information as well. |
190 |
> |
191 |
> ..... |
192 |
> |
193 |
> Also possibly of interest: |
194 |
> |
195 |
> The Gentoo Linux Genkernel Guide (outdated) |
196 |
> |
197 |
> Genkernel is a tool designed to help automate the kernel compilation |
198 |
> process. It'll help you create kernels similar to those on the |
199 |
> installation CDs. But while this document is still linked from several |
200 |
> of documents above, it's outdated and no longer maintained. As such, |
201 |
> while parts of it may be helpful, other parts may be more confusing than |
202 |
> helpful, unfortunately. As they say, YMMV. |
203 |
> |
204 |
> http://www.gentoo.org/doc/en/genkernel.xml |
205 |
> |
206 |
> |
207 |
> With those to help, and by copying the .config of your existing kernel |
208 |
> and using make oldconfig, you should be well on your way to a successful |
209 |
> kernel upgrade. A few of those and it'll be old hat. =:^) I'd probably |
210 |
> start with the general Compiling the Linux kernel link, to get a |
211 |
> background, then look at the upgrade guide to get a better idea of what |
212 |
> you're looking at here, then I'd use the config guide for configuration, |
213 |
> and if necessary, go back to the upgrade guide again to actually do it. |
214 |
> |
215 |
> Again, a very good place to start is with the .config from your existing |
216 |
> kernel, using make oldconfig to adapt it to the newer kernel. However, |
217 |
> it's reasonably likely that you'll still have questions about individual |
218 |
> choices. That's normal. You just experiment a bit until you get a |
219 |
> kernel that does what you need. As long as you don't remove your old |
220 |
> kernels, you can always reboot to one of them if the new kernel doesn't |
221 |
> work. And in fact, that's what even the Linux kernel hackers themselves |
222 |
> do. I do a lot of kernel testing here, and my rule of thumb is to keep |
223 |
> at least two known working kernels around, one good stable release from |
224 |
> the release BEFORE the one I'm testing, and a generally working current |
225 |
> kernel. Thus, until I have a known working current pre-release kernel, I |
226 |
> keep at least two previous release series kernels around, generally the |
227 |
> release and the first stable point release after it. (So for instance for |
228 |
> the current 2.6.30 series pre-release testing, I kept 2.6.29 and 2.6.29.1 |
229 |
> around, until I had a decently stable 2.6.30-rc2 plus version working to |
230 |
> my satisfaction, after which I deleted 2.6.29 but kept the latest stable |
231 |
> I had, 2.6.29.1, just in case something weird turned up with the 2.6.30- |
232 |
> rc2+ I had /thought/ was working.) |
233 |
> |
234 |
> For you, you'll probably keep 2.6.20.x around for some time, since you're |
235 |
> not going to be doing the kernel testing I do, and will presumably only |
236 |
> be running stable versions. |
237 |
> |
238 |
> ..... |
239 |
> |
240 |
> So anyway, between opening a portage bug and trying out a new kernel, one |
241 |
> way or another, you should eventually get around that bug. I'd encourage |
242 |
> you to keep working on both paths. In particular, if you do the clone |
243 |
> bug thing, even if you get a new kernel working that doesn't have the |
244 |
> issue, I'd encourage you to keep around your current 2.6.20 for testing |
245 |
> until that portage bug is resolved one way or another, because there's |
246 |
> probably others that will run into it as well. |
247 |
> |
248 |
|
249 |
|
250 |
-- |
251 |
John P. Burkett |
252 |
Department of Economics |
253 |
University of Rhode Island |
254 |
Kingston, RI 02881-0808 |
255 |
USA |
256 |
|
257 |
phone (401) 874-9195 |