Re: [Bug-wget] Features request

[Darshit Shah]

On 11/01, address@hidden wrote:
> Hi Darshit Shah ,
>
>
> Darshit Shah wrote:
> > Hi Taras,
> >
> > Thanks for your interest in Wget.
> >
> > On 10/31, address@hidden wrote:
> > > Hi all,
> > >
> > > I use wget, I think such features will be useful.Are there 
> > > objections against the below, or something is already implemented 
> > > and I did not find?
> > >
> > > 1) Maximal size of downloaded file (not all the batch) : do not 
> > > download if value in header present and stop and delete otherwise 
> > > if it is too big. I implemented it in my internal version. This 
> > > prevents downloading huge files when downloading from list and the 
> > > process is not controlled manually.
> > This might be an interesting option to consider. I'm not aware of 
> > any option to limit downloads based on a single file's size. 
> > However, you must remember that not all servers send a 
> > Content-Length header. In such scenarios, Wget *must* download the 
> > file before it realizes its actual size. In such a case, what would 
> > be the expected outcome? Should Wget delete the larger file? Or 
> > retain it?
> It should delete partial file.The full file must not be downloaded , 
> the download will stop when downloaded stuff reaches limit.For example 
> ,/maxsize 100000 while downloading 100 Mb file without size in header 
> - stops at 10 Mb and deletes the downloaded chunk.

I'm not too sure if this is feature is required. But if you already have it 
implemented, do share it on this list and we may consider merging it into the 
codebase.
> > > 2) Rename output to file001...bin if the URL produces junk name 
> > > (now unable to download).Additional option -  for any non-7bit 
> > > chars.
> > Could you please explain this issue further? What kind of junk name 
> > causes Wget to refuse downloading? A test case would be the best way 
> > to demonstrate this.
> HTTP download with compilcated URLs  like  
> http://someurl/foo.php?x=1?y=2jsdjdjdd. ,the wget produces file name 
> that cannot be created.I amnot sure,using not the last wget, 
> encountered such issues sometimes.In other cases  it produces 
> valid,but not readable name like 
> foo&fdjfdklfdfdfdl;&fkfkfdfkfkfkfk&bad.bin
> >
I still don't understand. I would suggest you give a concrete example with a 
real URL and the corresponding filename.

It seems like you're on Windows and Wget creates some filename which Windows 
refuses to create. We may only have to rename the output filenames for Windows 
in that case. Do provide a precise example so that we can understand the problem 
better.

> > > 3) -N should work with - O ,also for bad output names without -O.
> > Look at the manual page for Wget, under the --output-document 
> > section. It explains why -N and -O do not fit well together.
> This is also related to bad names. With -N -O the wget can look at 
> timestamp of file named as written in "-O". Please explain why this is 
> not OK.
I don't understand what you're asking for. Could you please be more clear on 
your position as to why -N and -O have a use-case when used together?
>
--- end quoted text ---

--
Thanking You,
Darshit Shah

pgpuuzy8sjB_E.pgp
Description: PGP signature