Since I'm not the only one that interpreted it that way, I've undeleted my CDPATH answer, though I've also upvoted Kusalananda's comments on the Q to clarify what exactly they're after. I won't hammer it closed, but I also notice no one else has VTC it yet, either.

@PrabhjotSingh You are welcome. Remember you can go up to 1K. I actually contributed to aprove one of your edits yesterday.

So slashdot does not let you login if you do not check a box they are allowed you "to contact" via email. Byebye slashdot account.

Could the same happen here someday?

don't you have email account set for that purpose ?

i.e. receiveing "contact and usefull news" (spelled Sierra Papa Alpha Mike)

Anyone able to help me sort out a filename-related issue in GNU Make? stackoverflow.com/questions/55471228

Suggestions for improvements also welcomed, of course.

It's amazing how quickly a question disappears from the frontpage on SO. Here on U&L, questions could stay for a day if the day is a bit slow. There, my GNU Make question got pushed down and off the page within 15 minutes.

@Archemar Out of principle, if memory does not fail me, I managed to give them a trow away email account,... However, the owner company nowadays seems to be just interested in spamming people

@Kusalananda And got an answer from the GNU Make maintainer (I think).

@FaheemMitha Aw dang. Nice! :-)

@Kusalananda Oh, you didn't notice? Yes, that's Paul Smith.

MadScientist?

@Kusalananda Yes, MadScientist.

I did not know.

I mean, he does not advertise his name on SO.

@Kusalananda Yes, he goes by the sobriquet of MadScientist.

I don't think his name is a secret as such.

I've got a feeling I've exchanged emails with him in the 1990s. But it's been a while.

@FaheemMitha What, really? MadScientist is the maintainer of GNU make?

Bombay is quite tiring. I was just outside for 2 1/2 hours, and I felt it. All that noise. All those people.

@terdon He is.

Actually, just follow the link on his profile page. I think he says so.

Well, the answer is solid enough for us to work with, so he can be whoever he wants to be :-)

Ah, different mad scientist. I thought you meant this one:

@terdon No, the one who answered @Kusalananda's question.

@FaheemMitha Yeah, figured it out. I assumed it was the other one because I knew that one :)

He's been GNU Make maintainer for a long time, I think.

But GNU maintainers don't exactly hog the limelight. So people mostly don't know who they are.

My memory was a bit off. I exchanged a couple of emails with him in June 2002. In the context of building some R extension, or something.

Well, shared library, not extension excactly.

Would anyone care to comment on dpi heuristics for scanning color documents or b&w documents? I.e. what dpi setting to use?

@FaheemMitha Also depends on what kind of quality is required; for text documents it probably doesn't need to be very high. Office scanners I've used the scanning speed gets extremely slow after certain DPI setting (so IMO practical to use as high DPI setting which still has acceptable scanning speed)

@sebasth Yes, it does depend. What do you use for text documents? color or grayscale, or b&w?

I have the Fijutsu ScanScap. Very popular, at least in the US.

for b/w documents (or documents which do not have any interesting/useful in color) I've used grayscale, usually been more practical should the scan be printed later (somehow some scanners do not get the paper white resulting in somewhat funny prints)

Wow; I simply don't have the brainpower to keep up with Brexit: "This bill would make it UK law for the PM to ask for an extension to prevent a no-deal, but it would be up to the EU to grant it - or not."

suppose it doesn't help that I haven't been keeping track of it at all -- diving into the middle of that article wasn't going to fix it :)

@sebasth Hmm, my scanner for some reason makes grayscale files much larger. Regardless, what dpi do you use?

For grayscale / color respectively?

@FaheemMitha Think the scanner I had was fastest at 300DPI (was quite old device)

@JeffSchaller I've mostly stopped trying. I wonder how many people are still keeping track.

@sebasth Ah. What model?

Can't recall, it was a scanner at work a couple years ago

(haven't had to scan much since)

@sebasth Oh? Hmm. I get to scan stuff all the time. They love paper here.

just to be printed out later, no doubt ;P

(larger file size could be related to file compression/optimization, you could try if something like jpegoptim has effect)

@sebasth Possibly.

For example, when I have to fill out a form these days, I tend to scan it and then write on top of it. Usually using LaTeX.

It's a bit laborious, but if I fill it in by hand, It's likely I will do something wrong.

huh, there is easy enough way to write on top of that scan using LaTeX?

@sebasth Well, I don't know how easy it is, but it's doable. I posted about it on SE somewhere. Possibly on U&L.

@sebasth unix.stackexchange.com/a/371629/4671

The grid is particularly useful. Otherwise you're guessing positioning.

This method is probably not so useful for non-TeX users, though.

yeah, but still something which sounds like more convenient to work with a GUI instead of measuring the document using a ruler

@sebasth If you have a grid, you don't need a ruler.

using the grid as ruler?

@sebasth Sure. Are you a LaTeX user?

I've been using it for awhile. Works quite well. All the GUIs I've used are buggy. Not fun. And TeX allows very precise handling. Which I like.

I use LaTeX for documents and sometimes for slides

@sebasth I use it for basically everything that I write. Other than emails.

And the occasional text file.

Still a beginner, though. And likely to remain so for the forseeable future.

At least there is tex.SE for us mere LaTeX mortals

@sebasth That's true. It's frequently a life saver.

I remember trying to write LaTeX on my own in the mid-90s. It was a terrifying experience. And much simpler than what I write now.

If I saw then the stuff that I write now, I'm sure I would have been quite intimidated.

@sebasth You don't have a TeX SE account, though.

ah, interesting

@sebasth So in summary, you'd go for around 300 dpi for grayscale? What about color?

Can I change color in grep --color-auto?

@sebasth What is?

@FaheemMitha haven't really had a need to do color scans, probably would experiment with same heuristic: quality good enough for reading while scanning being fast enough

(scan a few pages to test?)

@sebasth Speed isn't really the issue. Size is the issue.

@sebasth I've tested, but I don't have a color printer.

I imagine 300 DPI with color should be about as readable as with grayscale

Paul Smith on autoconf

I thought that we had a Linux monoculture these days. Politely ignoring the BSD folks, of course.

@sebasth I've been defaulting to 200 dpi recently. For color and b&w. My scanner calls the latter lineart. But I've been wondering (in my neurotic way) whether the color quality is good enough, since I can't actually print it. No color printer.

the only issue I've had has been if the software didn't adjust the colors to have background white, when printing he printer then prints slightly colored background too. didn't seem to get this issue with grayscale

@sebasth Is that with low dpi?

shouldn't be

Of course, not everyone cares about the color in color documents. And they shouldn't, because most of the time it's not functional.

I think it is some color correction automation which doens't work (for document scanning I would assume the software would adjust the colors to get white background)

@sebasth So not s low dpi issue, just a poor software issue?

yeah

And not SANE, I take it? SANE seems to work pretty well.

As far as I can tell, there are other layers in the mix.

it was some awful manufacturer software on Windows

I've been using gscan2pdf for a while. I suppose it must be using SANE at the backend, there isn't any other game in town.

@sebasth Ah

It's a bit buggy, but the author is very responsive. Sometimes takes him more than one try to fix a bug, though.

ME: Here's the correct way of doing it (describes in detail how to do it). And here's the quick and dirty hack that you will have issues with under some circumstances (describes hack and limitations). USER: Great! Thanks! The hack works, but I can't seem to get it to work under (specific conditions the hack won't work under).

@Kusalananda Welcome to the human race.

@Kusalananda this is why you don't tell USERs about hacks

You're a great moral support guys. ;-)

@sebasth I actually installed some Windows Virtualbox image (or similar) to check out the (proprietary) software for my scanner, which ran only under Windows, of course. Possibly OS X, too.

Anyway, I then discovered that SANE was actually better than that software. Which was a little disconcerting.

@FaheemMitha I can't believe you have been missing anything by not using their software. In my experience, manufacturers' (Windows) printer/scanner software is awful, to put it mildly

@sebasth It's been a while, but it seemed... not great. But the user reviews for the ScanSnap were stellar. And that includes the Windows software, because that's what people use, mostly. Windows, I mean.

Which is unfortunate, but such is life.

Perhaps secretly deep down, I still believe that proprietary software is better.

I mean, you have to pay for it. It must be better, right?

@FaheemMitha at least you might be assuring yourself about it ;D

@sebasth Hmm?

I just changed things up with my laptop's filesystem. I made a new data partition on my SDD, I've got it set to mount at boot via /etc/fstab, and I've got some bind mounts going on as well; namely from /partition/Documents to my $HOME/Documents/ and /partition/opt to /opt (I forgot to make /opt a separate partition when I installed, too much work to change it now)

I started up IntelliJ IDEA to work on a uni project, and it's saying stuff about slow file sync. I'm not sure why as it would be writing into $HOME/Documents, and the changes would automatically appear in the new data partition. Regardless, I found this site suggesting a fix: robbinespu.github.io/eng/2018/08/25/…

But I'm not sure what the consequences may be of increasing the number of watches. This question makes it sound like just a bit more memory utilization, which is only at 17.4% ATM

Anyone done this before, or able to clarify the consequences or explain how increasing the number of watches can increase syncing?

@Ungeheuer I take you're trying to share stuff across different operating systems? If not, perhaps take a moment to explain what you're trying to do.

And definitely mention that if you post a question.

In life, one often overlooks the obvious.

@FaheemMitha Yes, yes I am. I want Ubuntu to see if I can get more use out of my Nvidia card...under Fedora I can't even get an external monitor...dunno how anyone else makes this crap work

@Ungeheuer I don't follow.

I intend to share my data, e.g. documents, and programs e.g. IntelliJ via a data partition that is mounted at boot. A couple of bindfs mounts occur either at boot or on login to expose the appropriate parts of the partition to the appropriate places.

@Ungeheuer Ok. But what does your Nvidia card have to do with it?

Nothing. I was elaborating on my reason to install another Linux. Admittedly irrelevant.

@Ungeheuer Programs don't share so well. Binaries in general won't run under different Linuxes.

@Ungeheuer Ah, ok.

The only programs are IntelliJ, and Java SDK distributions

Everything else is documents or source code.

@Ungeheuer Oh.

@Ungeheuer Why do you need to copy in order to share?

That's just additional overhead.

The bind mounts are from a directory under data partition's mountpoint to the appropriate locations in the filesystem. I figured it was better than just symlinking.

Here's the question:unix.stackexchange.com/questions/510118/…

@Ungeheuer I've only a sketchy recollection of bind mounts, but clearly it involves syncing.

Is it well-written and understandable?

@FaheemMitha From toying with them now, any changes occurring on either end of the bind (the source or destination) is "immediately" reflected on the other end

@Ungeheuer Avoid putting text in images.

I have no idea how it's implemented.

@Ungeheuer I see.

@FaheemMitha I snipped the IntelliJ event notification, I'll give it some more space so the text around it is more distinct.

@Ungeheuer Are you actually seeing problems?

Not really, but if IntelliJ's complaining, then maybe it is an issue?

> Via the bind mount, this is actually located under the data partition's mountpoint. Or maybe it's also actually located under $HOME/Documents

Yea, that's the implementation detail I'm not certain about

I didn't read this carefully, but that's a little confusing.

@Ungeheuer Maybe figure it out first?

@terdon You'll enjoy this little April Fool's joke: bioexcel.xyz

@Kusalananda Oh, that's BRILLIANT! Thank you!

oh hey, I forgot to check in after the election. Congrats @Kusalananda :)

@FaheemMitha clarified a bit

@Seth Oh, thanks!

Personally, if all I was seeing was a warning, I'd be strongly tempted to ignore it till I saw actual symptoms. But that's me.

Also tangentially related:

@Seth No.

@FaheemMitha Tell that to a multitude of bioinformaticians (who thinks it is a database; hence the joke above).

Don't get me started. I hate spreadsheets. I also hate Microsoft. Take a guess how I feel about Excel.

@Seth Aaaaaargggggghhhhhhhh

@FaheemMitha Take a guess how other bioinformaticians feel when they have to process an Excel spreadsheet of multi-megabyte size on a Unix system :-)

@Kusalananda Unhappy? :-)

@FaheemMitha And frustrated. It's unfortunately quite common. That "BioExcel" joke page even links to a real publication outlining the issue with the names of genes, such as "SEPT2" and "MARCH1" that gets interpreted as dates by Excel and ends up as such in publications.

@Kusalananda I'm guessing you guys are speaking from first hand experience?

@FaheemMitha We take on a lot of various support requests from biomedical researchers and bioinformaticians all across Sweden, so yes.

@Kusalananda What, that's real?

> The spreadsheet software Microsoft Excel, when used with default settings, is known to convert gene names to dates and floating-point numbers. A programmatic scan of leading genomics journals reveals that approximately one-fifth of papers with supplementary Excel gene lists contain erroneous gene name conversions.

Mind blown.

Even after almost two decades working in this field. Mind blown.

@terdon Yes, that's definitely real.

@Kusalananda Come to think of it, I can't swear I've never done it myself. I know I've given lists of genes as supplementary excel files in at least one publication. But I'm pretty sure they were all uniprot names, not hugo, so they should be safe.

But wow.

:-)

OpenBSD just removed some classic X applications from the packages. Sigh, nostalgia. marc.info/?l=openbsd-ports-cvs&m=155421395702153&w=2

not xsnow! :(

@JeffSchaller "All of these were imported 20+ years ago and have been untouched since." Well, why touch perfect code?